cranfield university meletios pagonis electrical power ... · electrical power aspects of...
TRANSCRIPT
CRANFIELD UNIVERSITY
MELETIOS PAGONIS
ELECTRICAL POWER ASPECTS OF DISTRIBUTED
PROPULSION SYSTEMS IN TURBO-ELECTRIC POWERED
AIRCRAFT
SCHOOL OF AEROSPACE, TRANSPORT, AND
MANUFACTURING
PhD THESIS
Academic Year: 2012 - 2015
Supervisor: Professor Peter Malkin
October 2015
CRANFIELD UNIVERSITY
SCHOOL OF AEROSPACE TRANSPORT AND MANUFACTURING
Power And Propulsion Division
PhD THESIS
Academic Year 2012 - 2015
MELETIOS PAGONIS
Electrical Power Aspects of Distributed Propulsion Systems in
Turbo-electric Powered Aircraft
Supervisor: Professor Peter Malkin
October 2015
This thesis is submitted in partial fulfilment of the requirements for
the degree of Doctor of Philosophy
© Cranfield University 2015. All rights reserved. No part of this
publication may be reproduced without the written permission of the
copyright owner.
i
ABSTRACT
The aerospace industry is currently looking at options for fulfilling the
technological development targets set for the next aircraft generations.
Conventional engines and aircraft architectures are now at a maturity level
which makes the realisation of these targets extremely problematic. Radical
solutions seem to be necessary and Electric Distributed Propulsion is the most
promising concept for future aviation. Several studies showed that the viability
of this novel concept depends on the implementation of a superconducting
power network.
The particularities of a superconducting power network are described in this
study where novel components and new design conditions of these networks
are highlighted. Simulink models to estimate the weight of fully superconducting
machines have been developed in this research work producing a relatively
conservative prediction model compared to the NASA figures which are the only
reference available in the literature. A conceptual aircraft design architecture
implementing a superconducting secondary electrical power system is also
proposed. Depending on the size of the aircraft, and hence the electric load
demand, the proposed superconducting architecture proved to be up to three
times lighter than the current more electric configurations. The selection of such
a configuration will also align with the general tendency towards a
superconducting network for the proposed electric distributed propulsion
concept. In addition, the hybrid nature of these configurations has also been
explored and the potential enhanced role of energy storage mechanisms has
been further investigated leading to almost weight neutral but far more flexible
aircraft solutions. For the forecast timeframe battery technology seems the only
viable choice in terms of energy storage options. The anticipated weight of the
Lithium sulphur technology is the most promising for the proposed architectures
and for the timeframe under investigation. The whole study is based on
products and technologies which are expected to be available on the 2035
timeframe. However, future radical changes in energy storage technologies may
ii
be possible but the approach used in this study can be readily adapted to meet
such changes.
Keywords:
Superconductivity, electric, power, networks, machines, energy, storage, more,
electric, aircraft, battery
iii
ACKNOWLEDGEMENTS
First of all, I would like to express my gratitude to my supervisor, Professor
Peter Malkin, for his continuous guidance and support throughout these three
years. His passion for looking for new insights has been inspirational and his
contribution to my work has been crucial to the completion of this thesis.
Furthermore, I will always be grateful to Professor Pericles Pilidis for this
amazing opportunity he offered me when I most needed one. He believed in me
before I even believed in myself and for that I will always be indebted to him.
Through this PhD I had also the chance to work with numerous professionals
from Airbus Group Innovations (AGI) and Rolls Royce (RR) which also
significantly supported my work. Working along Graham Dodds and Frederick
Berg from AGI as well as with Mark Husband and John Cullen from RR has
been an amazing experience which helped me develop myself as a professional
engineer. In addition to them, I had also the pleasure to work and share
concerns and ambitions with several fellow Cranfield University students and
post graduates. Special regards to Joseph Palmer and Emanuele Pagone with
whom we worked countless hours together during the DEAP project.
I cannot begin to thank enough all these amazing people I met and befriended
during my Cranfield experience. Special thanks to my volleyball teammates who
have been a breath of fresh air during the stressful periods of my PhD.
Particularly to Giacomo, Radka, Giulio, Jakub, and Megane who have been the
best teammates I could have ever wished for! Moreover, my “Spanish group”
(i.e. Pedro, Ernest, Lola, Lelia, Alex, Paolo and Belen), thank you for all the fun
times you offered me. I would also never forget my “Bedfordians” and especially
my guardian angel Antonella for her constant support. Last but not least a
special thank you to all my loyal Greek friends that-no matter the distance-
remain important pieces of my life.
Finally, special thanks to my parents for their support and encouragement all
these years as well as to my brother and his wonderful family for being there for
me both at good and bad times.
iv
v
TABLE OF CONTENTS
ABSTRACT ......................................................................................................... i
ACKNOWLEDGEMENTS................................................................................... iii
TABLE OF CONTENTS ..................................................................................... v
LIST OF FIGURES ........................................................................................... viii
LIST OF TABLES ............................................................................................. xiv
LIST OF ABBREVIATIONS .............................................................................. xvi
1 Introduction & Project Specifications ............................................................... 1
1.1 Introduction ............................................................................................... 1
1.2 Future Goals and Trends in Aviation ......................................................... 2
1.2.1 Technology goals for next aircraft generations ................................... 3
1.2.2 Potential future design options for the civil aerospace industry .......... 4
1.3 DEAP Project ............................................................................................ 8
1.4 Thesis methodology and structure .......................................................... 10
2 Literature Review .......................................................................................... 12
2.1 Distributed Propulsion (DP) .................................................................... 12
2.1.1 Small Gas Turbines (GTs) Concept ................................................. 13
2.1.2 Distributed Driven Fans .................................................................... 17
2.1.3 Electric Distributed Propulsion with a Conventional Electric Power
Network ..................................................................................................... 20
2.1.4 N3-X Turbo-electric Distributed Propulsion Configuration ................ 22
2.1.5 Distributed Electrical Aerospace Propulsion European Projects ...... 24
2.1.6 Distributed Propulsion Summary ...................................................... 26
2.2 Superconductivity.................................................................................... 27
2.2.1 Superconducting Materials ............................................................... 28
2.2.2 Superconducting Components ......................................................... 32
2.3 Cooling system ....................................................................................... 35
2.3.1 Cryogenic Fluid with a Heat Sink ..................................................... 35
2.3.2 Cryo-coolers ..................................................................................... 37
2.4 Summary ................................................................................................ 40
3 Design of Autonomous Electrical Power Networks ........................................ 42
3.1 Introduction to Electric Power Network Design ....................................... 42
3.2 Conventional Design of Autonomous Electric Power Networks (EPNs) .. 43
3.2.1 Proposed Autonomous Power Network Design Process .................. 43
3.2.2 Hybrid/electric ship design process example ................................... 50
3.3 Superconducting Electric Power Network Elements ............................... 57
3.3.1 Superconducting Electrical Machines............................................... 57
3.3.2 Superconducting Switches ............................................................... 58
3.3.3 Superconducting Fault Current Limiters (SFCLs) ............................. 58
3.3.4 Protection System and Converters ................................................... 62
3.3.5 Cooling System ................................................................................ 64
vi
3.4 Superconducting Electric Power Networks Design and Operation ......... 64
3.4.1 Basic Parameters Selection ............................................................. 64
3.4.2 Current splitting ................................................................................ 69
3.4.3 Electro-magnetic Forces .................................................................. 69
3.5 Summary ................................................................................................ 70
4 Superconducting Electrical Machines ............................................................ 72
4.1 Status and State of the Art ...................................................................... 72
4.1.1 Superconducting Synchronous Machines ........................................ 72
4.1.2 Homopolar DC Superconducting Machines ..................................... 78
4.1.3 Superconducting Induction Machines............................................... 79
4.1.4 Programmable Superconducting AC Machine (PSAM) Project ........ 80
4.1.5 Summary .......................................................................................... 81
4.2 Weight Estimation of Fully Superconducting Machines .......................... 85
4.2.1 Torque per unit of rotor volume (TRV) method ................................. 85
4.2.2 Relationship between rotor and stator dimensions ........................... 87
4.2.3 Basic Assumptions ........................................................................... 89
4.2.4 Models Description ........................................................................... 91
4.3 Sensitivity Study ..................................................................................... 96
4.3.1 The environmental screen ................................................................ 97
4.3.2 TRV Factor ..................................................................................... 102
4.3.3 Active Power Density ..................................................................... 104
4.3.4 Cryostat Weight .............................................................................. 105
4.3.5 Winding factor ................................................................................ 106
4.4 Key Model Limitations ........................................................................... 107
4.5 Model Validation ................................................................................... 109
5 Superconducting Electric Aircraft (SEA) ...................................................... 111
5.1 More Electric Aircraft (MEA) Concept ................................................... 111
5.1.1 MEA Concept Description .............................................................. 111
5.1.2 Airbus 380 ...................................................................................... 114
5.1.3 Boeing “Dreamliner” 787 ................................................................ 115
5.1.4 Going Beyond 787: Challenges and design options ....................... 117
5.2 Superconducting Electric Aircraft Approach .......................................... 119
5.2.1 787 Electrical System Overview ..................................................... 119
5.2.2 Superconducting Version of 787 Electrical Power Network ............ 124
5.3 MEA and SEA Weight and Efficiency comparison studies (based on
the Boeing 787 aircraft) ............................................................................... 129
5.3.1 Basic Assumptions ......................................................................... 130
5.3.2 Results and Comments .................................................................. 135
5.4 SEA Sensitivity/Scalability Studies........................................................ 140
5.4.1 Reference Aircraft Description ....................................................... 141
5.4.2 MEA and SEA Simulink models overview ...................................... 144
5.4.3 Weight Trends in reference aircraft ................................................ 147
vii
5.4.4 Final Remarks ................................................................................ 155
5.5 Key Study Limitations ........................................................................... 157
6 Novel Flight Cycles for Hybrid/Electric Aircraft Using Energy Storage ........ 160
6.1 Energy Storage ..................................................................................... 161
6.1.1 Batteries ......................................................................................... 161
6.1.2 Supercapacitors ............................................................................. 167
6.1.3 Superconducting Magnetic Energy Storage (SMES) ..................... 170
6.2 Novel Hybrid Configurations and Flight Cycles ..................................... 171
6.2.1 Baseline Aircraft and Mission Profile .............................................. 172
6.2.2 Overview of the Modelling Approach .............................................. 174
6.2.3 HEDP proposed configurations ...................................................... 176
6.2.4 Final Remarks ................................................................................ 192
6.3 Sensitivity Study for Hybrid Configurations for aircraft of different
sizes............................................................................................................ 193
6.3.1 Reference Aircraft Mission Profiles ................................................ 194
6.3.2 Results and comments ................................................................... 197
6.3.3 Final Remarks ................................................................................ 207
6.4 Key study Limitations ............................................................................ 210
6.5 Roadmap for Novel Flight Cycles Investigation .................................... 212
7 Conclusions and Future Work ..................................................................... 215
7.1 Concluding Remarks ............................................................................. 215
7.1.1 Superconducting Power Networks (SPNs) ..................................... 216
7.1.2 Superconducting Electrical Machines............................................. 217
7.1.3 Superconducting Electric Aircraft (SEA) ......................................... 217
7.1.4 Novel Flight Cycles with Energy Storage ....................................... 218
7.1.5 Key Findings Summary .................................................................. 218
7.2 Recommendations for future work ........................................................ 219
REFERENCES ............................................................................................... 224
APPENDICES ................................................................................................ 239
viii
LIST OF FIGURES
Figure 1 Electric Distributed Propulsion System ................................................. 2
Figure 2 Technology Development S-Curve (Scocco, 2006) ............................. 3
Figure 3 Advanced Single-Aisle Aircraft configuration with rear-mounted open rotor engines (Guynn D. et al., 2011) ......................................................... 6
Figure 4 SUGAR Volt: Boeing’s proposed design for next aircraft generations (courtesy of Boeing) .................................................................................... 7
Figure 5 NASA N3-X Hybrid wing body aircraft with TeDP (Kim et al., 2013) .... 8
Figure 6 Front and planform view of the proposed DEAP aircraft baseline (Alderman, 2014) ......................................................................................... 9
Figure 7 Distributed Propulsion Concepts Historical Overview (Gohardani, Doulgeris and Singh, 2011) ....................................................................... 13
Figure 8 Wing box savings with different number of engines (Eggenspieler, 2006) ......................................................................................................... 15
Figure 9 Electric Propulsion System (Luongo et al., 2009) ............................... 19
Figure 10 Dual-use commercial/military transport vehicle (Green, Schiltgen and Gibson, 2012) ............................................................................................ 20
Figure 11 Hybrid electric distributed propulsion system example (Schiltgen et al., 2012) .................................................................................................... 21
Figure 12 N3-X Hybrid Wing Body Aircraft Turbo-electric Distributed Propulsion Concept (Felder, Kim and Brown, 2009).................................................... 22
Figure 13 EADS Innovations Work E-Thrust Concept Configuration (Courtesy of Airbus) ....................................................................................................... 25
Figure 14 Basic electric architecture case in DEAP project .............................. 26
Figure 15 Critical T-H-I Diagram for a superconducting material (www.what-when-how.com, 2015) ............................................................................... 28
Figure 16 Micrograph showing the cross-section of an as-drawn BSCCO wire (courtesy of Applied Superconductivity Research Center)......................... 30
Figure 17 2nd Generation 𝑴𝒈𝑩𝟐 wires improved current density. (Courtesy of Hyper Tech Research Columbus) ............................................................. 31
Figure 18 Typical HTS Cable structure (Courtesy of Suptech.com) ................. 32
Figure 19 An example of a LH2 power system TeDP configuration (Masson et al., 2007) .................................................................................................... 36
Figure 20 Reverse Brayton Cryo-cooler study (Berg et al., 2015a) .................. 38
ix
Figure 21 Projected development of cryo-coolers optimised for aerospace applications (Palmer, Pagonis and Malkin, 2015) ...................................... 39
Figure 22 Design Process diagram of a conventional power network (Malkin and Pagonis, 2013) ................................................................................... 44
Figure 23 Synthesis of a waveform from harmonics ......................................... 48
Figure 24 Dynamic phenomena with their corresponding timescales in a power network: A. Electro-magnetic transients, B. Synchronous machine transients, C. Quasi steady state, and D. Steady-state phenomena (Andersson, 2006) ..................................................................................... 49
Figure 25 Electric/Hybrid Ship Propulsion System Diagram (Malkin and Pagonis, 2014) .......................................................................................... 51
Figure 26 Diesel-electric ship propulsion plant (marine.man.eu, 2015) ............ 52
Figure 27 Bus Voltage Levels for given total required power demand (Doerry and Fireman, 2006) ................................................................................... 54
Figure 28 Typical losses diagram of a hybrid-electric ship propulsion system . 56
Figure 29 Simulink model of a single phase system with SFCL ....................... 59
Figure 30 Simulink model of SFCL subsystem ................................................. 61
Figure 31 Single phase current waveforms in a system with and without a SFCL .................................................................................................................. 61
Figure 32 TeDP Protection System Proposed Architecture (Armstrong et al., 2012) ......................................................................................................... 63
Figure 33 Illustration of Paschen’s Law (Paschen, 1889) ................................. 66
Figure 34 Comparison between the transmission losses of a conventional and a superconducting cable (Masuda et al., 2004) ............................................ 68
Figure 35 Typical losses diagram of a propulsion system using a superconducting network ........................................................................... 69
Figure 36 25MW 120 RPM superconducting synchronous motor U.S Navy conceptual design (Gamble et al., 2002) ................................................... 74
Figure 37 Siemens HTS Synchronous Machine Test Bed (image courtesy of Siemens) ................................................................................................... 75
Figure 38 HTS Motor using Gd-Ba-Cu-O bulk magnets schematic illustration (Matsuzaki et al., 2005) ............................................................................. 76
Figure 39 The first fully superconducting motor (Takeda, Oota and Togawa, 2006) ......................................................................................................... 76
Figure 40 200 kW HTS Reluctance Motor (Oswald et al., 2005) ...................... 77
x
Figure 41 Layout University of Southampton’s 100 kW HTS machine (Wen et al., 2009) .................................................................................................... 78
Figure 42 HTS DC Homopolar Motor (image courtesy of General Atomics). ... 79
Figure 43 Schematic diagram of the test system of a fabricated HTS induction motor installed in a metal cryostat (Nakamura et al., 2006) ....................... 80
Figure 44 PSAM Machine Arrangement (Berg and Dodds, 2013) .................... 81
Figure 45 Weight vs Torque of singly superconducting machines .................... 84
Figure 46 Rotor vs. stator dimensions relationship graph ................................ 88
Figure 47 General view of the DEAP superconducting electrical machine (Courtesy of the DEAP project) ................................................................. 90
Figure 48 Simulink model (first version) for the weight estimation of fully superconducting electrical machines ......................................................... 93
Figure 49 Simulink model (second version) for the weight estimation of fully superconducting electrical machines ......................................................... 95
Figure 50 Mean stator to outer stator subsystem ............................................. 96
Figure 51 Environmental Screen Simulink Sub-model ..................................... 99
Figure 52 Overall weight of a superconducting machine (a) without environscreen, (b) with iron screen (4-poles machine) and (c) with iron screen (8-poles machine). ....................................................................... 100
Figure 53 Overall weight of a superconducting machine (a) without environscreen, (b) with aluminium screen (4-poles machine) and (c) with aluminium screen (8-poles machine). ...................................................... 101
Figure 54 TRV Factor Vs. Total Weight of Fully Superconducting Machines . 103
Figure 55 Active Power Density Vs. Total Weight of Fully Superconducting Machines ................................................................................................. 105
Figure 56 Cryostat Weight Factor Vs. Total Weight of Fully Superconducting Machines ................................................................................................. 106
Figure 57 Winding Factor Vs. Total Weight of Fully Superconducting Machines ................................................................................................................ 107
Figure 58 Conventional secondary power systems (Jones, 2002) ................. 112
Figure 59 Comparison between conventional and MEA systems (Provost, 2002) ................................................................................................................ 113
Figure 60 Airbus 380 aircraft (image courtesy of Airbus) ............................... 114
Figure 61 A380 Power distribution system (Abdel-Fadil, Eid and Abdel-Salam, 2013) ....................................................................................................... 115
xi
Figure 62 Boeing Dreamliner 787 Aircraft (image courtesy of Boeing) ........... 116
Figure 63 787’s electrical system compared to traditional aircraft architecture ................................................................................................................ 117
Figure 64 Total Electrical Power Demand during several flight stages of the 787 aircraft (Whyatt and Chick, 2012) ............................................................ 120
Figure 65 Electric loads and efficiencies diagram of the 787 electrical power network .................................................................................................... 121
Figure 66 Trent 1000 three shaft configuration (Ojha and Raghava, 2014) ... 122
Figure 67 Variable Frequency Starter Generator (VFSG) used in 787 (Clark, 2012) ....................................................................................................... 123
Figure 68 Electrical Power Distribution System in 787 (Moir and Seabridge, 2013) ....................................................................................................... 124
Figure 69 Boeing 787 and Airbus A350 size (www.AviationExplorer.com, 2015) ................................................................................................................ 127
Figure 70 Tested behaviour of power electronic devices at cryogenic temperatures (Leong, 2011) .................................................................... 129
Figure 71 Weight per meter of conventional copper cable with PVC insulation ................................................................................................................ 133
Figure 72 Weight Comparison between different 787 MEA and SEA configurations .......................................................................................... 136
Figure 73 Electric loads and efficiencies diagram of the 787 electrical power network in SEA case (DEAP estimates) .................................................. 138
Figure 74 Electric loads and efficiencies diagram of the 787 electrical power network in SEA case (NASA estimates) .................................................. 139
Figure 75 MEA’s Electric Power Network Simulink Model .............................. 146
Figure 76 SEA’s Electric Power Network Simulink Model (Superconducting DEAP case) ............................................................................................. 147
Figure 77 VFSGs’ weight for each reference aircraft in all four different versions ................................................................................................................ 148
Figure 78 Power Electronics’ Simulink Model................................................. 150
Figure 79 Power Electronics’ weight for each reference aircraft in all four different versions ..................................................................................... 150
Figure 80 Cables’ weight for each reference aircraft in all four different versions ................................................................................................................ 153
Figure 81 Cooling system’s weight for each reference aircraft in all four different versions ................................................................................................... 154
xii
Figure 82 Electrical Power Network total weight for each reference aircraft in all four different versions .............................................................................. 155
Figure 83 Electrical Power Network total weight for different electric load requirements ............................................................................................ 156
Figure 84 Current Specific Energy values of different battery types ............... 162
Figure 85 Typical specific energy values for different battery technologies (batteryuniversity.com, 2015) .................................................................. 164
Figure 86 Status of Li-S batteries compared to the United States Advanced Battery Consortium (USABC) baseline standards (Mikhaylik et al., 2015) ................................................................................................................ 166
Figure 87 Energy and power density of different energy storage options (Hampton, 2013) ...................................................................................... 169
Figure 88 Schematic of a SMES device (Molina, 2010) ................................. 170
Figure 89 Mission profile of the DEAP aircraft ................................................ 173
Figure 90 Weight vs. Shaft Power of Turboshaft/turboprop engines .............. 175
Figure 91 HEDP case 1 energy storage sizing Simulink model ...................... 177
Figure 92 Case 1 GTA (red line) and Energy Storage (blue line) power output with time .................................................................................................. 178
Figure 93 Case 1 Energy Storage State of Charge (SoC) in kJ with time (s) . 179
Figure 94 Energy storage weight estimation Simulink model ......................... 180
Figure 95 Case 1 Electric Components sizing Simulink model....................... 182
Figure 96 HEDP case 2 energy storage sizing Simulink model ...................... 184
Figure 97 Case 2 GTAs (red line) and Energy Storage (blue line) power output (kW) with time (s) ..................................................................................... 185
Figure 98 Case 2 Energy Storage State of Charge (SoC) in kJ with time (s) . 186
Figure 99 Weight of the battery system vs. specific energy assumptions for Case 2 configuration ................................................................................ 188
Figure 100 Case 3 GTA 1 (red line), GTA 2 (green line) and Energy Storage (blue line) power output (kW) with time (s) .............................................. 190
Figure 101 Case 3 Energy Storage State of Charge (SoC) in kJ with time (s) 190
Figure 102 Overall Weight comparison for the different cases ....................... 193
Figure 103 Main engines’ thrust ratings vs number of passengers in the reference aircraft ..................................................................................... 196
Figure 104 Mission Profiles of the reference aircraft ...................................... 197
xiii
Figure 105 Weight Comparison between Case 1, Case 3, and a configuration without energy storage for all the reference aircraft ................................. 208
Figure 106 Li-sulphur weight vs. reference aircraft SLS power requirements 209
Figure 107 HEDP Architecture Proposal ........................................................ 213
Figure_A-1 Schematic diagram showing RBC next to the Simulink model .... 241
xiv
LIST OF TABLES
Table 1 NASA and ACARE goals for next aircraft generations .......................... 4
Table 2 Summary of the main characteristics of several superconducting materials .................................................................................................... 32
Table 3 Industry wire performance requirements for various device applications (Larbalestier et al., 2001)* ......................................................................... 33
Table 4 Diesel-electric propulsion plant main parameters ................................ 53
Table 5 Diesel-electric propulsion plant switchboard parameters ..................... 55
Table 6 Fundamental parameters of a resistive SFCL ..................................... 60
Table 7 List of singly superconducting electrical machines .............................. 83
Table 8 Inputs/Outputs of the Simulink models for the weight estimation of fully superconducting machines ........................................................................ 92
Table 9 Initial assumed values for the model’s inputs ...................................... 97
Table 10 Inputs/Outputs of the Environmental Screen Subsystem .................. 98
Table 11 Comparison between NASA and TRV model weight estimates ...... 109
Table 12 VFSGs key variables values for each case ..................................... 131
Table 13 Power electronics key variables values for each case..................... 132
Table 14 Main Cable line key variables values for each case ........................ 134
Table 15 BOEING 737-900 Main characteristics ............................................ 142
Table 16 BOEING 777-300 Main characteristics ............................................ 142
Table 17 A350 Main characteristics ............................................................... 143
Table 18 A380 Main characteristics ............................................................... 143
Table 19 CRJ-1000 Main characteristics ........................................................ 144
Table 20 Main inputs and outputs of MEA’s power network Simulink model .. 145
Table 21 Weight per meter (kg/m) of the main transmission lines for each reference aircraft in all four different versions .......................................... 152
Table 22 Comparison of different types of battery currently in use (www.batteryspace.com, 2015) ............................................................... 162
Table 23 Battery technology summary and sizing parameters ....................... 167
Table 24 Main characteristics of DEAP Aircraft .............................................. 172
Table 25 Case 1 GTA and Energy Storage sizing factors .............................. 179
xv
Table 26 Case 1 Energy storage technology sizing values ............................ 181
Table 27 Summary of Case 1 components’ weight ........................................ 182
Table 28 Case 2 GTA and Energy Storage sizing factors .............................. 184
Table 29 Case 2 Energy storage technology sizing values ............................ 186
Table 30 Summary of Case 2 components’ weight ........................................ 187
Table 31 Case 3 GTA and Energy Storage sizing factors .............................. 189
Table 32 Case 3 Energy storage technology sizing values ............................ 191
Table 33 Summary of Case 3 components’ weight ........................................ 192
Table 34 Reference Aircraft Main Characteristics .......................................... 195
Table 35 CRJ-100’s Energy storage technology sizing values ....................... 198
Table 36 Summary of CRJ-100 components’ weight ..................................... 199
Table 37 B737‘s Energy storage technology sizing values ............................ 200
Table 38 Summary of B737 components’ weight ........................................... 201
Table 39 B787‘s Energy storage technology sizing values ............................ 202
Table 40 Summary of B787 components’ weight ........................................... 203
Table 41 A350‘s Energy storage technology sizing values ............................ 204
Table 42 Summary of A350 components’ weight ........................................... 205
Table 43 A380‘s Energy storage technology sizing values ............................ 206
Table 44 Summary of A380 components’ weight ........................................... 206
Table_A-1 Conventional Electrical machines dimensions .............................. 239
Table_A-2 Main inputs/outputs of Cryo-cooler Simulink models .................... 242
Table_A-3 List of turboshaft/turboprop engines.............................................. 244
xvi
LIST OF ABBREVIATIONS
Nomenclature
AC
ACARE
ACT
AGI
APU
ATRU
BLAC
BLI
BSCCO
BWB
CAA
CAEP
CB
CFD
CNT
CO2
CVG
DAPRA
DC
DEAP
DoD
DP
ECS
EDLC
EOR
EPN
E.S.
FAA
FOD
FC
GE
GT
GTA
HEDP
HPC
HTS
Alternating Current
Advisory Council for Aviation Research and Innovation in Europe
Advanced Capacitors Technology
Airbus Group Innovations
Auxiliary Power Unit
Auto Transformer Rectifier Unit
Brushless AC machines
Boundary Layer Ingestion
Bismuth Strontium Calcium Copper Oxide
Blended Wing Body
Civil Aviation Authority
Committee on Aviation Environmental Protection
Circuit Breaker
Computational Fluid Dynamics
Carbon Nanotubes
Carbon Dioxide
Constant Velocity Gearbox
Defence Advanced Research Projects Agency
Direct Current
Distributed Electrical Aerospace Propulsion
Depth-of-Discharge
Distributed Propulsion
Environmental Control System
Electric Double Layer Capacitor
End-Of-Runway
Electric Power Network
Energy Storage
Federal Aviation Administration
Foreign Object Damage
Fuel Cell
General Electric
Gas Turbine
Gas Turbine Alternator
Hybrid Electric Distributed Propulsion
High Power Compressor
High Temperature Superconductors
xvii
HVLC
IDG
IM
IP
KERI
LED
LH2
LIC
LP
LTS
LVHC
MEA
MEE
MEL
MgB2
NASA
NOx
PAX
P.E.
PM
PMAD
PSAM
PVC
RMS
RPDU
RPM
RR
SEA
SFCL
SIG
SLS
SM
SMES
SoC
SOTA
SPN
SPS
SPU
SR
High Voltage Low Current
Integrated Drive Generator
Induction Motor
Intermediate Power
Korea Electro-technology Research Institute
Light-Emitting Diode
Liquid Hydrogen
Lithium-Ion Capacitors
Low Pressure
Low Temperature Superconductors
Low Voltage High Current
More Electric Aircraft
More Electric Engine
Maximum engine Electric Loading
Magnesium Diboride
National Aeronautics and Space Administration
Nitrogen Oxide
Passengers
Power Electronics
Permanent Magnet
Power Management and Distribution
Programmable Superconducting AC Machine Project
Polyvinyl Chloride
Root Mean Square
Remote Power Distribution Unit
Rotations Per Minute
Rolls Royce
Superconducting Electric Aircraft
Superconducting Fault Current Limiter
Superconducting Induction Generator
Sea Level Standard
Superconducting Machine
Superconducting Magnetic Energy Storage
State of Charge
State-Of-The-Art
Superconducting Power Network
Secondary Power System
Secondary Power Unit
Switched Reluctance
xviii
SUGAR
TeDP
TO
TOC
TRL
TRU
TRV
TSB
TSFC
UAV
VF
VFSG
WAI
WF
YBCO
Subsonic Ultra Green Aircraft Research
Turbo-electric Distributed Propulsion
Take-off
Top-of-Climb
Technology Readiness Level
Transformer Rectifier Unit
Torque per unit of Rotor Volume
Technology Strategy Board
Thrust Specific Fuel Consumption
Unmanned Aerial Vehicle
Variable Frequency
Variable Frequency Starter Generator
Wing Anti-Icing
Wound Field
Yttrium Barium Copper Oxide
Symbols
B Magnetic field
H* Irreversibility Field
Hc Upper Critical Field
Jc Critical Current Density
𝑚𝑐𝑟𝑦𝑜 Mass of the cryo-cooler
nprop Propulsive efficiency
𝑃𝑖𝑛 Input Power requirement
Tc Critical Temperature
vj Jet velocity
vo Free-stream velocity
𝑉𝑆 System’s Voltage
𝐼𝑆 System’s Current
F Frequency
𝑖𝑠 Steady state alternating current
𝑖𝑡 Transient direct current
U Voltage
xix
Z Impedance
R Resistance
L Inductance
𝜔 Fundamental frequency
T Time
θ Voltage angle
𝑖𝑚𝑚 Maximum momentary short circuit
𝑃𝑆ℎ𝑎𝑓𝑡 Shaft Propulsion Power
𝑛𝑡𝑟𝑎𝑛𝑠 Electric transmission efficiency
𝑃𝐵𝑝𝑟𝑜𝑝 Engine brake power for transmission
𝑃𝑒𝑙𝑒𝑐 Electric consumer load
𝑛𝑎𝑙𝑡 Alternator efficiency
𝑃𝐵𝑒𝑙𝑒𝑐 Engine brake power for consumer
𝑃𝐸 Total engine brake power demand
𝑃𝑇𝑜𝑡𝑎𝑙 Total engine brake power installed
𝐼𝐺𝑠𝑐 Generator short circuit current
𝑛 Number of generators/motors
𝑃𝐺𝑒𝑛 Rated power of the generator
𝑉𝑟 Rated Voltage
𝑥𝑑" Sub-transient reactance
cos 𝜑 Power factor
𝐼𝑀𝑠𝑐 Motor short circuit current
𝑃𝑚𝑜𝑡 Rated power of the motor
𝑉𝐵 Breakdown Voltage
𝑝 Atmospheric pressure
𝑑 Distance
𝐹𝑒𝑚 Electromagnetic forces per unit length
𝜇0 Permeability constant
P Electromagnetic Power
T Torque
𝜔𝑚 Rotational speed
N Number of rotations
A Electric loading
xx
B Magnetic loading
𝑚 Number of phases
𝑇𝑝ℎ Number of turns in series per phase
𝐷 Diameter
𝛷 Fundamental flux/pole
𝑝 Number of pair poles
𝐿𝑠𝑡𝑘 Stack length
𝐸 Electro-magnetic force (emf)
𝑘𝑤 Winding factor
𝑣𝑟 Rotor volume
𝐿 Length
𝑉𝑝ℎ Phase Voltage
𝐼𝑝ℎ RMS phase current
𝑡𝑐 Thickness of the environmental screen
𝛾 Environmental screen density
𝑟𝑠 Mean stator radius
𝑟𝑥 Inner screen radius
𝑊𝑠𝑐𝑟𝑒𝑒𝑛 Environmental screen mass
LiCoO2 Lithium cobalt oxide
LiFePO4 Lithium iron phosphate
NiMH Nickel Metal Hydride
Li-ion Lithium ion
NMC Nickel-manganece-cobalt
Li-air Lithium air
Li-S Lithium sulphur
1
1 Introduction & Project Specifications
1.1 Introduction
The scope of this research study is to investigate, mainly in terms of weight,
several system architectures of novel future aircraft designs which are
characterised by their multiple propulsion power sources (“distributed
propulsion”). This will include a more/all electric approach and the presence of a
superconducting power network.
The concept of distributed propulsion (DP) for aerospace applications implies
the separation and distribution of the propulsive system that allows
improvements in “propulsive efficiencies” to enhance the overall vehicle
performance. It has been shown that electric DP is the most beneficial
configuration, hence the more electric approach described in this thesis.
However, the feasibility of this approach depends on the weight reduction of the
whole system. In order to achieve the latter a superconducting network might be
necessary.
A more graphical representation of the proposed design can be seen in figure 1.
In this graph a couple of Gas Turbines Alternators (GTAs) are responsible for
providing the required propulsive power to a number of motor driven fans, whilst
the energy storage system is also operating as a supplementary power source.
This figure gives only a preliminary idea of the concept being investigated in this
research work. Several variations of this approach will be described later in this
thesis.
2
Figure 1 Electric Distributed Propulsion System
It is important to clarify the reasons behind the investigation of such a disruptive
technology. The following subchapter will present the expected improvements
of next aircraft generations and it will demonstrate the incapability of
conventional configurations to satisfy these demands.
1.2 Future Goals and Trends in Aviation
The aerospace industry is a sector that throughout the recent past has
consistently made significant improvements in regards to the performance and
fuel efficiency of civil aircraft. The last 50 years or so (i.e. after the introduction
of the jet engine) the majority of researchers have been focusing on the
implementation of technology advances regarding the propulsion system,
airframe alternatives, materials selection, aerodynamics etc. with the goal of a
more efficient and safer aircraft. However, one important thing to note is that
recently the rate of improvement has been significantly decreased as a result of
the physical limitations that have been reached in many technologies. Figure 2
illustrates a typical technology development s-curve which includes the
introduction, expansion and maturation of innovations that most industries,
including aerospace, experience.
3
Figure 2 Technology Development S-Curve (Scocco, 2006)
Aerospace industry has been approaching the phase where any performance
improvements take longer to be attained and require significantly more effort
and money. Combining the latter with the testing emission targets that important
institutions such as NASA (National Aeronautics and Space Administration) and
ACARE (Advisory Council for Aviation Research and Innovation in Europe)
have set for the future aviation underlines the importance of innovative aircraft
design options that will open numerous unexplored paths for designing civil and
military airplanes.
1.2.1 Technology goals for next aircraft generations
The expected continued growth in air traffic, combined with the increased
demand for minimisation of environmental damage in all aspects of technology
has led the aviation industry to search for ways to diminish the negative impact
of future aircraft on environment. NASA released some very aggressive targets
for next generation commercial airplanes (Ashcraft W. et al., 2011),
concentrating on basic key aspects like noise, NOx emissions, and last but not
least fuel burn (CO2 emissions). In 2001 ACARE also set some similar targets
(Graham R., Hall A. and Morales V., 2014). Both these goals have been
summarised in table 1. Note that different references have been chosen in
these two cases. More specifically, ACARE used as a reference a 2000 aircraft
for all three categories while NASA used a 2005 state-of-the-art aircraft for the
fuel burn savings, an aircraft-engine NOx emissions standard being set by the
Committee on Aviation Environmental Protection (CAEP) and a cumulative sum
4
of lateral, flyover and approach noise certification points under the FAA stage 4
noise regulation.
Table 1 NASA and ACARE goals for next aircraft generations
Category
NASA ACARE
N+2
(~2020)
N+3
(~2035)
Vision 2020
FlightPath 2050
Fuel Burn -50% -60%* -50% -75%
Noise -42dB -71dB -50% -65%
NOx Emissions -75% -80%* -80% -90%
*Note that these values have changed throughout the years and different numbers can be found
in literature, however the most recent ones were chosen.
It is clear that such optimistic targets will not be achieved following the
“conventional approach”. For this reason all the important aerospace companies
and research centres are looking for rather disruptive technologies that could
significantly alter the aviation industry. These concepts will be briefly described
in the next subchapter while a more detailed description of the chosen approach
will be presented in Chapter 2.
1.2.2 Potential future design options for the civil aerospace industry
Several different concepts have been suggested and developed by various
organisations for the next aircraft generations, some of which will be briefly
described here. Naturally, the so-called Turbo-electric Distributed Propulsion
concept is the approach which will be investigated in more depth in this
research study as it is considered by the author as the most promising
technology.
Open rotor concept
One of the concepts that have attracted significant interest lately is the
unducted -“open rotor”- propulsion approach. This is not a new concept at all,
but it was firstly investigated in the late 1970s and early 80s triggered by the
sharp increase of the fuel prices during this period. More specifically, NASA’s
Advanced Turboprop Project was one of the most high-profile projects of this
5
era claiming fuel cost benefits of almost 50% (Whitlow B. and Sievers K.,
1988). However, the fuel benefits did not come without a penalty. Noise levels
being achieved with this configuration were below FAA stage 3 limits, whilst the
probability of foreign object damage (FOD) was also significantly increased.
Note that all these results (positive and negative) were based on the state-of-
the-art technology of that era and do not reflect the current situation. Since this
concept was out of any agenda for almost two decades efforts to re-establish
the know-how were necessary. An initial assessment of the open rotor
propulsion concept capabilities as well as predicted benefits was carried out by
NASA (Guynn D. et al., 2011) focusing on counter-rotating pusher approach
with a rear-mounted installation (Figure 3). The first results showed Thrust
Specific Fuel Consumption (TSFC) reductions of around 30% in the top-of-climb
(TOC) phase and more than 45% for the Sea Level Standard (SLS) and Take-
off (TO) phase. Even more optimistic results about the NOx emissions were
presented, claiming a reduction of the order of 80%. On the other hand, the
Thrust-to-Weight ratio of the open rotor engines was more than 15% lower than
other turbofan technologies mitigating the fuel savings of this configuration.
Finally, the results of the noise measurements were not so disheartening but
were presented for demonstration purposes only and their reliability at this point
is questionable.
6
Figure 3 Advanced Single-Aisle Aircraft configuration with rear-mounted open
rotor engines (Guynn D. et al., 2011)
Boeing Subsonic Ultra Green Aircraft Research (“SUGAR” )
Boeing Corporation has done a lot of work aiming at fulfilling NASA’s targets for
future aviation and to that direction a series of aircraft designs were proposed.
The “SUGAR” family (acronym for Subsonic Ultra Green Aircraft Research)
consists of five different aircraft designs: SUGAR Free, Refined SUGAR,
SUGAR High, SUGAR Ray, and SUGAR Volt. The latter was the one that
presented results closer to the desired ones. More specifically, fuel burn
savings greater than 70% and large emission reductions could be achieved.
SUGAR Volt is characterised by its hybrid electric-gas turbine engines designed
by General Electric (GE) and the use of batteries during take-off and landing.
Boeing SUGAR team concluded that hybrid electric energy technology is the
clear winner for future aviation and has the greatest potential of achieving
NASA’s targets (Stephenson, 2010). However, in order this concept to become
feasible important improvements on battery technologies are necessary.
Several modifications of SUGAR Volt design will be investigated in this project
thesis (Chapter 6).
7
Figure 4 SUGAR Volt: Boeing’s proposed design for next aircraft generations
(courtesy of Boeing)
NASA N3-X Turbo-electric Distributed Propulsion Concept (TeDP)
One of the most high profile studies that was the basis for a lot more to come is
the so-called N3-X model. This configuration consists of two turboshaft engines
driving two superconducting electrical generators. The primary function of these
devices is to produce electrical power, rather than thrust. These two turbo-
generators are mounted on the wing-tips, a location that proved to be more
beneficial for such a model. The electrical power is transmitted along redundant
superconducting electrical cables to an array of propulsors embedded in the
entire upper trailing edge of the fuselage section of the aircraft. In the N3-X
configuration there are 14 propulsors, each with a superconducting motor driven
fan. The aforementioned configuration can be seen in Figure 5.
8
Figure 5 NASA N3-X Hybrid wing body aircraft with TeDP (Kim et al., 2013)
The initial results showed a 70-72% mission fuel burn reduction compared to a
B777-200LR-like vehicle (Felder et al., 2011b). A more extended analysis of the
N3-X project will follow in the next chapter (Chapter 2) where all the
technologies, architectures, enablers and limitations of this concept will be
described.
Distributed Electrical Aerospace Propulsion (DEAP) Project
This is a UK government funded project with the participation of Airbus Group
Innovations (AGI), Rolls Royce (RR), and Cranfield University. It is also linked
with this research study where parts of it were inputs for the several phases of
this project. Thus, a more detailed description of this project and its links with
this study is necessary and will follow in the next subchapter (1.3).
1.3 DEAP Project
Airbus and Rolls Royce are lately exploring different paths for the propulsion
system of future aircraft. In order to achieve this, they joined forces in the DEAP
project also collaborating with Cranfield University. This project investigates key
innovative technologies that will enable improved fuel economy and reduced
emissions for future airliner designs having Distributed Propulsion (DP) and
Boundary Layer Ingestion (BLI). The main objectives of the project were to
9
evaluate distributed propulsion concepts as well as to analyse the feasibility and
potentials of a more electric approach combined with BLI. A concept plane
proposed by Airbus was the baseline for any models and calculations being
carried out (Figure 6).
Figure 6 Front and planform view of the proposed DEAP aircraft baseline
(Alderman, 2014)
The project consists of three main work packages. The first one is about the
aircraft integration study and investigates a number of techniques to define the
initial aircraft and to review other candidate configurations as part of exploitation
case studies. The second word package-which is also related to this thesis
study-had as an initial objective the development of an electrical system model
to a high fidelity, the consideration of different transmission solutions and the
contribution to the optimised fan design. Finally, the third package concentrated
on the development and testing of the BLI fan design.
Some of this research study results were also used and verified during the
DEAP project. The challenges of a possible fully superconducting network were
pointed out, whilst the weight estimation of components that have not been built
yet was an important input to the DEAP project. Energy storage possibilities
were also investigated in this thesis report supporting the work package 2 of
DEAP project. A more detailed description of the structure of this thesis will
follow in the next section.
10
1.4 Thesis methodology and structure
The overall aim of the study is to assess the challenges, limitations and
potential benefits of the TeDP concept as a future aircraft propulsion approach.
Early studies have shown superconductivity to be one of the main enablers of
this concept; hence primarily a focus on the design of such a network was
necessary. Some of the components of this network have never been used
before in airborne applications, whilst others have not even been built yet.
Hence, it was important to investigate the performance of these components,
the interaction between them and some crucial attributes such as their weight
which became the number one priority as the research was moving forward.
The novelty of the whole approach however creates numerous design
possibilities for the hybrid electric aircraft under investigation. Some of these
possibilities were explored during this study enhancing the attractiveness of the
concept. The thesis outline could be summarised as follows:
A literature review will follow in Chapter 2. This review will initially cover
the DP concept comparing the possible modifications of this approach.
All the previous and current studies will be presented whilst the electric
hybrid approach will be emphasized. Superconductivity as a
phenomenon will be presented and a brief description of the cooling
options in this type of aircraft will also be described.
The study of Autonomous Power Networks will be the main focus of
the third Chapter. Firstly the design process of a conventional power
network will be described, whilst a specific working example will also be
presented. After that, the novel components of a superconducting
network will be analysed and a comparison between the design process
of a conventional and a superconducting network will conclude this
section.
Chapter 4 describes a method to estimate the weight of fully
superconducting machines. Corresponding machine models will be
demonstrated and used throughout this research study. Sensitivity
studies in regards to the several inputs of the models will conclude this
section.
11
A Superconducting Electric Aircraft approach will be the subject of
Chapter 5. In this approach a superconducting version of the More
Electric Aircraft (MEA) concept will be described testing the limits and the
design possibilities that such an approach could create.
A novel flight cycle study will then be reviewed in Chapter 6. This
chapter will focus on the numerous design options of the DEAP concept
for the optimisation of the propulsion system of such an electric hybrid
aircraft. The role of energy storage subsystems will also be explored.
Batteries, Supercapacitors, and Superconducting Magnetic Energy
Storage (SMES) will be investigated as possible energy storage options
for the TeDP concept.
The results and outcomes of the study will be discussed in Chapter 7
(Conclusions). A summary of the most important concluding remarks
and key findings of this research study will be pointed out. Finally, future
work will be the last section of this study. Since this project applies for
the timeframe 2035+, this Chapter will reasonably be an important
contribution guiding the future research studies on the fields and
knowledge gaps of this innovative concept.
12
2 Literature Review
Before each chapter of this research study the corresponding literature review
will be described. An overview of the main aspects of the novel concept under
investigation will be presented in the current section of this thesis aiming on
determining the optimal solution for the next aircraft generations and revealing
the knowledge gaps of the proposed design.
Initially the historical evolution of the Distributed Propulsion (DP) concept will be
presented and different modifications of DP configurations will be compared.
The most promising of them (i.e. Turboelectric Distributed Propulsion) will be
further analysed by both using a more conventional electrical power network as
well as a superconducting version of it. This will lead to a quick overview of the
phenomenon of superconductivity focusing on the materials and actual
applications being used. Finally, a look on the possible cooling system options
will conclude the first part of this study covering up all the background work that
led to the implementation of this research study.
2.1 Distributed Propulsion (DP)
A first definition of the distributed propulsion term has been given in Chapter
1.1. However, DP could also more simplistically be described as any propulsion
system which spreads the thrust requirements along the span of the aircraft.
This could happen either by spreading the engines’ exhaust or by spreading the
propulsors. As a concept DP is not new at all with the first conceptual designs
having started in the early twenties. More specifically, in 1924 Manzel proposed
a patent for the propulsion system of airships, aircraft and like, consisted of
multiple propeller units arranged in two rows (Manzel, 1924). This patent was
aiming on an airship capable to ascent without a special landing field. Griffith in
1954 (Griffith, 1954) proposed a configuration with one master turbofan and
several slave turbojet units distributed spanwise. The motivation behind this
concept was the possibility of thrust vectoring and short take-off and landing
phases. The next studies of distributed propulsion were driven by the lower
weight to thrust ratio of the small gas turbines but the main breakthrough started
13
in the mid-70s where the fuel cost started to rise sharply and alternative
propulsion options, including DP, attracted more interest. Figure 7 highlights
some of the conceptual milestones in regards to the DP concept.
Figure 7 Distributed Propulsion Concepts Historical Overview (Gohardani,
Doulgeris and Singh, 2011)
2.1.1 Small Gas Turbines (GTs) Concept
The majority of the initial DP studies were focusing on the distribution of smaller
gas turbines. Although this concept presented some benefits, it has also shown
some detrimental effects that were quite difficult to be overcome. In regards to
the former, breaking the propulsion system into smaller units could lead to a
more flexible and robust system, which can be characterised by its multi-
functionality and possibly lower weight. Another benefit could be the so-called
economies of scale, because by increasing the number of engines in an aircraft
mass manufacturing might be necessary (Ameyogo, 2007). Moreover, by
14
distributing the engine weight and thrust loads a lighter wing structure could
become feasible. The possibility of Boundary Layer Ingestion (BLI) is also
enhanced in such a configuration, since a larger number of small engines will
occupy more space in the wingspan, increasing the area available for boundary
layer ingestion. BLI potential positive effect in future aircraft was one of the main
areas of interest of both the DEAP project as well as for many other studies
carried out in Cranfield University (Liu, 2013). Finally, noise reduction could be
another advantage of the small gas turbines DP approach.
Theoretically, GT engine weight should be proportional to the cube of its
dimensions. This could potentially lead in a weight benefit for the distributed
configurations. However, as we scale down the engines, a number of
component dimension constraints and manufacturing limitations can be
observed. Due to the lack of research in small gas turbines and in new
materials technology, considerable weight gains can be feasible only at the
expense of severe performance degradation (Ameyogo, 2007).
The number and weight of the engines also affect the overall airframe weight.
According to Torenbeek (Torenbeek, 1992) a single engine per wing already
brings an approximately 3.5% reduction in wing structural weight. However,
Eggenspieler (Eggenspieler, 2006) claimed that the maximum reduction in
bending moment is not achieved with the maximum number of engines, but with
a number of 11 engines per wing. The lack of space to optimise engine
placement in the wing seems to be the main reason that further increase in the
number of engines per wing does not lead to further wing box weight savings.
The root moment reduction in the case of 11 engines per wing approaches
13%.
15
Figure 8 Wing box savings with different number of engines (Eggenspieler, 2006)
Notwithstanding the potential benefits of such a configuration, there are also
many important disadvantages mainly derived from the small gas turbines
themselves. Performance wise, small engines are less efficient since they suffer
from manufacturing limitations that can make the cooling procedure impossible
(Harada, 2003). Combustion efficiency also is degraded as combustors are
scaled down. Moreover, tip losses will be relatively large for small engines, as
the gap between the casing and the blade is proportionally wider, a fact that has
a direct effect on compressor performance (Schaub, Vlasic and Moustapha,
1993). Additionally, below their critical value Reynolds numbers can lead to
laminar flow and early separation, decreasing that way the components’
efficiency.
The reliability of the engines is also an important issue for any aircraft
configuration. Clearly, the more the engines the more possible is one of them to
fail. However, thrust losses will be significantly less even if more than one
engine fail simultaneously. On the other hand, if we place the engines close to
each other, the possibility of cascading engines failure increases, since a rotor
fragment could hit one of the adjacent engines. A chain reaction could therefore
eliminate half of the available engines, jeopardising airframe reliability. In
80%
82%
84%
86%
88%
90%
92%
94%
96%
98%
100%
1 2 5 11 21 39 58
We
igh
t /
Bas
elin
e W
eig
ht
Number of Engines
Final Relative Primary Wing Box Weight
16
addition, the large number of engines is directly connected with the number of
sensors which are necessary to control the whole propulsion system. Thus,
controls and instrumentation seem to be a significant barrier in the case of
distributed propulsion configurations. If preventive maintenance is used, less
instrumentation would be necessary, but even then the cost of instrumentation
will still be high.
Another significant issue for this configuration is the location of the engines.
Conventional airframes do not seem to be able to exploit the benefits of the
distributed propulsion. Embedded engine configurations seem to present the
best alternative. By embedding engines in the wing wave drag rise could be
avoided, nacelle drag could be reduced and BLI would be enabled. The
available space for the engines in the wing is the main concern in such a
configuration. The engines that are now used are too large to embed them into
a conventional wing. On the other hand, there is a limit of how much the size of
the engine could be reduced, since the core cannot be too small. For the above
reasons airframes such as the Blended Wing Body (BWB) might present a more
optimal solution for DP concepts.
It becomes clear that current technology does not allow optimism in regards to
the feasibility of small gas turbine distributed propulsion. Fuel consumption
appears to be the most important hurdle, as well as the effect of nacelle drag
which aggravates the fuel consumption issue even further. Furthermore, main
disadvantages could be considered the size effects that lead to reduced overall
thermal efficiency (Laskaridis et al., 2015). However, in the long term there are
technologies that with improvements could make this project possible. In
general, for aerodynamic reasons small gas turbine compressors have lower
overall pressure ratios. The use of lighter heat exchangers with a higher
effectiveness is therefore critical. The use of high temperature materials, such
as ceramic matrix composites could, up to a point, eliminate the problem of
manufacturing small-scale turbine blade cooling systems (Williams et al., 2008).
The most critical enabler for such a configuration will be the alternative
airframes. Although BWB may not seem an ideal solution, many research
17
studies showed a lot of potential benefits using such an airframe in a DP
configuration (Ko et al., 2003).
2.1.2 Distributed Driven Fans
Distributed driven fans are a more attractive case of distributed propulsion for
many reasons. This configuration includes a core that generates power for a
number of fans, transmitted via a transmission line. The distributed driven fans
concept not only has most of the benefits that were previously mentioned for
small gas turbine distributed propulsion, but also yields a significantly better
overall efficiency of the system (Ameyogo and Singh, 2007). Clearly, this makes
them more attractive from an environmental point of view. The three main
components of this propulsion system are: the core, the transmission system
and the fans.
Whilst high thermal efficiencies and fan propulsive efficiencies would be
relatively easy to achieve by using either current or advanced technologies, the
transmission of the power from the core to the fans will be a challenge and will
most probably need unconventional technologies. Buquet (Buquet, 2007)
carried out a research study of three different potential transmission systems for
a distributed propulsion configuration: mechanical, core gas, and electrical
transmission.
Mechanical transmission: This type of transmission forms the smallest
technological step of all the options. A free power turbine is connected to
the core gas turbine and provides power to the fans through an
arrangement of gearboxes and shafts. This configuration is similar to the
one being used in helicopters. The biggest drawback of this method is
clearly the weight which could become prohibitive as the number of fans
rises. Buquet (Buquet, 2007) concluded that with current materials, the
shaft weight required to transfer power to 20 fans would exceed
30,000lbs, in addition to the 20,000lbs weight of the gearboxes. This
method is therefore prohibitive for a distribution system consisting of
more than two fans.
18
Gas transmission: In this case, instead of using a free power turbine the
gas turbine exhaust is mounted into a plenum chamber. This chamber
feeds a series of tip turbines through a manifold. These tip turbines then
drive the distributed fans. Although the insulation required to transmit the
core gas without excessive heat losses is relatively light, the weight of
the ducts, of the supporting structure and of the tip turbines is still
excessive. Moreover, the issue of ducting hot gas through the wing
structure will create significant problems for an aerospace application
(Kim, 2010). Furthermore, unknowns such as tip turbine size limitations
and weight make this transmission method less attractive.
Electrical transmission: Although this kind of transmission is the least
developed, it shows the greatest potential to achieve low-weight, high
efficiency energy transmission. Extensive analysis of this concept will
follow in the upcoming sections.
The first advantage of the electrical transmission method is safety. Since high-
energy rotating shafts or hot pressurized ducts are not involved, the system is
safer. Also, if one of the fans failed it would be easier to isolate the fault without
disturbing the normal function of the propulsion system. Furthermore,
transmission efficiency in electric power systems is at least comparable to
simple gearbox systems and significantly higher than gas transmission systems.
However, perhaps the greatest benefit of electrical distribution is the flexibility
that offers. If electric motors were mounted before each of the distributed fan, it
will allow them to be independently driven at different rotational speeds (Figure
9). Each generator will be mechanically linked with a core engine, while the fans
will be electrically connected to the generators through a kind of “electrical
gearbox”. Thus, since generators can tolerate higher speeds than fans, the LP
shafts of the core engines could run faster, decreasing the number of stages
and the weight of the turboshaft engines. Moreover, decoupling torque and
speed would lead to more control flexibility, enabling a better trade-off between
design and off-design performance. Additionally, the use of electrical power
transmission allows a high degree of freedom to place the generators and fan
19
modules to their most advantageous location. Power transmission lines do not
require a particular strong and heavy supporting structure.
Figure 9 Electric Propulsion System (Luongo et al., 2009)
Although conventional technology does not allow optimism about the feasibility
of distributed driven fans propulsion systems due to the weight penalties
associated with the electrical machines, there are still ways to improve the
attractiveness of the whole concept. Superconducting electrical machines and
networks have been proposed by many researchers as the main enabler of the
DP concept. Superconductivity could lead both to higher overall efficiencies and
to lighter electric machines and components. The lighter the transmission
system, the higher the optimum number of fans will be. The phenomenon of
superconductivity will be described in this chapter (Chapter 2.2), whilst the
benefits and constraints of a fully superconducting network will be presented in
Chapter 3 and an extensive analysis of superconducting electrical machines will
follow in Chapter 4. Superconducting networks require a cooling system to
operate and a number of cooling methods will be presented later in this
literature review study.
20
2.1.3 Electric Distributed Propulsion with a Conventional Electric
Power Network
The first reasonable step towards the direction of hybrid electric distributed
propulsion was the investigation of an aircraft implementing a conventional
electric power network. The feasibility, potential benefits and possible hurdles of
such a configuration could be easier explored by using electric components that
are currently used in several industries and their technology is mature enough
to be used in such a sensitive application as the aerospace. NASA, as part of a
program called ‘N+2 Distributed Propulsion Studies’, developed the idea of a
dual-use commercial/military transport vehicle, an aircraft with a rather
conventional design, with two turbo-generators mounted in the middle of the
wings and 16 propulsive fans each directly driven by an electric motor (Figure
10).
Figure 10 Dual-use commercial/military transport vehicle (Green, Schiltgen and
Gibson, 2012)
In the aircraft above conventional electric machines and cables were used with
an operating temperature around 450K. Green et al.(Green, Schiltgen and
Gibson, 2012) created a program in Matlab aiming to analyse hybrid propulsion
systems for future aircraft concepts and managed to make some useful
conclusions about the feasibility of NASA’s aircraft. For this study a single
design point at 3000 ft. altitude and Mach number of 0.65 was chosen for all the
components of the propulsion system. However, in such a configuration where
21
there is no mechanical decoupling between the various components different
design points might prove more beneficial.
Figure 11 Hybrid electric distributed propulsion system example (Schiltgen et al.,
2012)
Figure 11 demonstrates the system architecture that has been investigated
during this NASA study. The turboshaft engine produces power to drive the
generator, which produces the required electrical power to the whole network.
This power is then transmitted through a transformer to a central wing box
controller which directs this power to each motor, as well as to several other on-
board systems. Another series of controllers and transformers are mounted
before every motor-fan couple where the necessary thrust is being produced.
The study concluded that using conventional electric components (i.e. motors,
generators, cables, and controllers) the propulsion system becomes much
heavier than the state of the art conventional propulsion systems. However,
benefits such as decoupled energy management, increased thermal and
propulsive efficiency leading to possible reduced fuel consumption and
capability of the aircraft to “fly smart”, are facts that cannot be ignored. More
specifically, this aircraft design showed a potential 40% reduction in fuel burn,
providing the implementation of an embedded turbo-electric distributed
22
propulsion system combined with cryogenic cooling (Schiltgen et al., 2011).
Ways to diminish the weight penalties and to take advantage as much as
possible the benefits of such a configuration need to be found. With this in mind
NASA developed the N3-X model as part of the N+3 Advanced Aircraft
Concepts.
2.1.4 N3-X Turbo-electric Distributed Propulsion Configuration
NASA is the organisation which has investigated in more depth the concept of
DP. After several research studies the most suitable configuration of distributed
propulsion seems to be the so-called N3-X configuration (Figure 12).
Figure 12 N3-X Hybrid Wing Body Aircraft Turbo-electric Distributed Propulsion
Concept (Felder, Kim and Brown, 2009)
This configuration consists of two turboshaft engines driving two
superconducting electrical generators. The primary function of these
components is to produce electrical power and not thrust as in the conventional
architectures. The electrical power is transmitted along redundant
superconducting electrical cables to an array of propulsors embedded in the
entire upper trailing edge of the fuselage section of the aircraft. In the N3-X
configuration there are 14 propulsors, each consisted of a superconducting
motor driven fan.
23
The two turbo-generators are assumed to be located on the wing-tips. Although
this is not a common location for turbo-machinery in this application it offers
some advantages. First of all, it minimises the risks to the aircraft and the
passengers in the event of a turbine disk failure. Moreover, it allows the inlet to
ingest free-stream air. As most of the energy of the gas stream is extracted by
the power turbine in order to drive the generator, the exhaust velocity is low
which consequently should result in low jet noise; hence one of the main goals
of future aviation becomes easier to be achieved. Some bending moment relief
in the normal direction is also possible due to this location (Felder, Kim and
Brown, 2009). Additionally, research conducted by NASA in 1970 showed
reductions in induced drag of up to 40% when a device that produces thrust is
located at the wing-tip; this is due to the higher velocity thrust stream which
reduces the wing-tip vortex well downstream of the wing itself. Last but not least
the phenomenon of BLI will also be facilitated. The force required to decelerate
the incoming air is the diffusion (or inlet drag) of the propulsion system and is
proportional to the velocity of the incoming air. Thus, the propulsor inlet flow is
decelerated by upper fuselage surface viscous forces and allows the propulsor
system to take advantage of the wake, by reducing the inlet velocity of the
propulsor and hence reducing the amount of inlet drag (Chengyuan et al.,
2012). Furthermore, if the fan nozzle is not choked, the slower the inlet velocity
the slower the exit velocity will be. The propulsive efficiency is given by the
following equation:
nprop =2
1+vj
vo
(2-1)
Where vo is the free-stream velocity and vj is the nozzle exit velocity. This
equation is valid only if the inlet velocity equals the free-stream velocity (Pilidis,
2012). However, ingesting the boundary layer can result in significant losses.
The trade-off between the benefits and drawbacks of BLI is a complicated task,
but in order to enhance the potential benefits a hybrid wing body aircraft should
have the following characteristics (Felder et al., 2011b):
24
Inlets that ingest a large percentage of the upper surface boundary layer
Inlets that are located near the trailing edge
Continuous inlets and nozzles for minimizing external wetted area
No boundary ingestion through the core engines, so that thermal
efficiency will not be affected
Minimum number of core engines, and
High transmission efficiency.
Boundary layer ingestion is an important feature of the distributed propulsion
concept. Many researchers have investigated several aspects of this
phenomenon, but a lot of research is still to be done in this field. NASA (Felder
et al., 2011b) examined the effect of boundary layer ingestion on turboelectric
distributed propulsion systems in more detail, while Cranfield University
(Doulgeris et al., 2012) investigated the dynamic response and high cycle
fatigue analysis of fan blades under inlet distortion, a phenomenon linked with
the boundary layer ingestion. Finally, several MSc students of Cranfield
University have investigated the effect of boundary layer ingestion in areas such
as ducted axial fans (Costi, 2012), noise generation (Chambon, 2012) etc. and
studies concerning this subject are still under progress(Valencia and Nalianda,
2015).
2.1.5 Distributed Electrical Aerospace Propulsion European Projects
NASA is not the only organisation investigating the TeDP concept. There are
also some European projects mainly leaded by Airbus which explore both the
distributed propulsion and the hybrid/electric concepts for the next generation
aircraft. The baseline for the current and future projects is the Airbus’ E-Thrust
configuration (Figure 13).
25
Figure 13 EADS Innovations Work E-Thrust Concept Configuration (Courtesy of
Airbus)
This distributed propulsion configuration consists of a single large turbine
engine producing electricity to six ducted fans which provide the required thrust.
This concept presents many similarities with NASA’s N3-X models. Engines are
optimised in turning fuel to shaft power; fans are optimised for higher propulsive
efficiency, whilst both superconductivity and BLI are also taken into
consideration. More specifically, the gas turbine is embedded in the tail so that it
ingests the fuselage boundary layer and the motor driven by the engine is
assumed to be superconducting. Finally, advanced lithium ion batteries are
used mainly as a supplementary power unit during take-off and climb, whilst
they are recharged during cruise with power provided by the engine. At all times
the batteries should have sufficient energy to power the aircraft in the case of
the turbine failure (Warwick, 2013).
The DEAP project which was earlier described (1.3) and constitutes an
important part of this thesis was basically based on the E-Thrust project. In the
DEAP configuration there are two gas turbine generators producing the power
that drives eight motor driven fans who produce the require thrust (Berg et al.,
2015a). Both the electrical machines and the transmission system are
superconducting, whilst different architectures in regards to AC and DC
distribution were investigated. Figure 6 demonstrates the proposed airframe,
26
while Figure 14 presents an example of the basic electric architecture
investigated during the DEAP project.
Figure 14 Basic electric architecture case in DEAP project
Boundary layer ingestion was also one of the main fields of study for the DEAP
configuration. The DP propulsors were placed in the boundary layer at the rear
of the fuselage. CFD studies were carried out to determine the characteristics of
the BLI intake and the possible efficiency penalties on the fans derived from the
fuselage BLI (Wright et al., 2015).
In the following chapters, DEAP project will often be mentioned as a reference
since significant work of this research study was carried out under the Cranfield
University required deliverables during this two-year TSB funded program.
2.1.6 Distributed Propulsion Summary
Although TeDP seems as a very promising design concept, there are some
particular aspects of it that should be further investigated in order the proposed
configuration to be successful. As it has already been mentioned, the major
barrier of this design is the weight penalty of the conventional electrical
machines.
Adding almost 14,000 lbs of motors, generators, power electronics and
transmission equipment instead of a single shaft and gearbox may seem
27
impossible to result any fuel savings (Felder, Tong and Chu, 2012). But the
comparison is more complicated than that. Turbo-electric powered aircraft
saves weight by eliminating the gearbox, the pylons and most importantly
reduces the fuel load (better efficiency) which leads to a lighter propulsive
system; hence a lighter aircraft.
Superconductivity is perhaps the main enabler of manufacturing lighter
machines, but there are a lot of side effects which someone needs to take into
consideration. A major challenge will be the manufacturing of superconducting
motors and generators with superconducting filaments of sufficiently small
diameter to keep losses low in the stator. The phenomenon of superconductivity
will be described in more detail in the following subchapter.
2.2 Superconductivity
Superconductivity is a phenomenon observed in certain materials that present
true zero electrical resistance and expulsion of magnetic fields when cooled
below a critical temperature. As a phenomenon it was discovered in 1911 by
Dutch physicist Heike Kamerlingh Onnes. Mercury was the first material which
Onnes discovered suddenly losing its resistivity when cooled to the temperature
of liquid Helium (i.e. 4K). In the next few decades the same behaviour was
observed in several metals, alloys, and compounds. However, strong interest in
the field of superconductivity was mainly revived in the 80’s when the first so-
called High Temperature Superconducting (HTS) materials were discovered
(Malkin and Pagonis, 2013). Critical temperatures of up to 110 K have been
recently reached, whilst significant efforts to fully understand the capabilities
and limits of these materials have been made. In addition to temperature, a
superconducting material should not exceed certain limits of current density and
magnetic field (Figure 15).
28
Figure 15 Critical T-H-I Diagram for a superconducting material (www.what-when-
how.com, 2015)
2.2.1 Superconducting Materials
There are numerous materials presenting the superconducting properties but
not all of them are appropriate for an industrial or an aerospace power
application. The requirement of cooling to near liquid helium temperatures was
the main limiting factor of applying the superconducting technology to any
application for several decades. However, the discovery of superconducting
materials at temperatures above 77K initiated a new era in the field of
superconductivity. For power applications similar to the ones investigated in this
research study there are three main superconducting materials which have
attracted more interest; Bismuth Strontium Calcium Copper Oxide (BSCCO),
Yttrium Barium Copper Oxide (YBCO), and Magnesium Diboride (𝑀𝑔𝐵2).
YBCO is the first material ever discovered to become superconducting above
the boiling point of liquid nitrogen (i.e. 77K) having a critical temperature over
90K. It was discovered in 1986 by Georg Bednorz and Alex Mueller who were
working in IBM Switzerland (Bellis, 2015) and it was the base of many HTS
materials to come. The significant breakthrough of the discovery of YBCO is the
much lower cost of the refrigerant used to cool the material below their critical
temperature. YBCO has been used as the main material in the rotor of several
superconducting electrical motor and generators prototypes (Chapter 4.1), as
well as in electrical equipment such as Superconducting Magnetic Energy
29
Storage (SMES) devices (Jin et al., 2014) and Superconducting Fault Current
Limiters (SFCLs) (Kim, Sim and Hyun, 2006). Both these superconducting
applications will be analysed in following sections of this study. There are two
primary reasons that prevent the use of these materials in several applications.
Firstly, their low critical current densities compared to the competitive
superconducting materials and secondly the difficulties in processing these
materials into the commonly required wire form. The critical current density for
an YBCO material operating at a temperature of 77K can be found around 350
𝐴/𝑐𝑚2 (Sözeri, Özkan and Ghazanfari, 2007).
BSCCO on the other hand was first discovered around 1987 by H. Maeda and
his colleagues at the National Research Institute for Metals in Japan (Maeda et
al., 1988). Its high critical temperature of above 105K attracted the interest of
many industries the years following its discovery. The understanding of this
material even today is a complicated task. However, the last decade BSCCO
has been progressed to the point of being commercially available with solid
mechanical properties in operational temperatures around 77K (Scanlan,
Malozemoff and Larbalestier, 2004). Many superconducting electrical machines
prototypes have been using BSCOO for their rotor (Chapter 4.1), whilst other
electrical equipment such as HTS transformers of maximum power rating
between 500 kVA and 1MVA have been using BSCOO-2223 winding cooled at
77K (McConnell, Walker and Mehta, 2000). Furthermore, BSCOO filaments
have also been used in SMES devices (Shi et al., 2007).
30
BSCCO was also the first HTS material that was used for making practical
superconducting wires. It is typically available in tape form (Figure 16) making
the production of BSCOO material in wire form a challenging task. The critical
current density (Jc) of BSCOO-2223 based wires has been the research focus
of many companies for over 20 years steadily improving its performance over
time. Long-length wires (>150m) with maximum current up to 170 A (Scanlan,
Malozemoff and Larbalestier, 2004) and critical current density of 12-14 𝑘𝐴/𝑐𝑚2
at 77 K self-field (Jin et al., 2014) are nowadays available.
Figure 16 Micrograph showing the cross-section of an as-drawn BSCCO wire
(courtesy of Applied Superconductivity Research Center)
Finally, 𝑀𝑔𝐵2 is the simplest and less expensive superconducting material
under investigation. It was discovered in 2001 by the group of Akimitsu
(Akimitsu, 2001) and is considered as the “conventional superconductor” with
the highest critical temperature (i.e. 39K). Although its critical temperature is
lower than the one of the HTS materials, its simple and robust mechanical
properties make it an attractive option for many applications. It is also available
in fine twisted filaments and in a wire form reducing the AC losses that the other
two popular superconducting materials (i.e. YBCO and BSCOO) present. For
this reason it seems reasonable to believe that future fully superconducting
machines will use 𝑀𝑔𝐵2 as their main material for their stator (more information
in Chapter 4). Furthermore, its relatively low cost and its capability of very sharp
transition for the superconducting to the normal state enhance their
attractiveness for protection devices such as SFCLs (Shcherbakov, 2011).
Finally, recent publications about 𝑀𝑔𝐵2 indicate that further developments of
31
this material will be available with even higher critical current densities 𝐽𝑐 (figure
below) (Li et al., 2012).
Figure 17 2nd Generation 𝑴𝒈𝑩𝟐 wires improved current density. (Courtesy of Hyper
Tech Research Columbus)
Table 2 summarises some of the critical values of the aforementioned
superconducting materials. Generally, it is really difficult to give absolute values
for these parameters since they are a function of several secondary factors. In
the following table the current density was based on an operational temperature
of 77K for both the YBCO and BSCOO case (Scanlan, Malozemoff and
Larbalestier, 2004), whilst the value for the 𝑀𝑔𝐵2 case was based on an
operational temperature of 20K. The last column is independent of the
operational temperature and it is just an indication of the upper 𝐻𝑐 limit in each
material. Note that although BSCOO has a higher 𝐻𝑐 limit it has a much lower
irreversibility field 𝐻∗ than YBCO which enhance YBCO’s use in creating high
field magnets (Golovashkin et al., 1991). The value of 74 T in the 𝑀𝑔𝐵2 case
can be reached only in thin films.
32
Table 2 Summary of the main characteristics of several superconducting
materials
Material Critical
Temperature (K)
Upper Critical
Field (T)
Critical Current Density
(𝒌𝑨/𝒄𝒎𝟐)
YBCO 93 170 0.35
BSCOO 110 200 12-14
𝑀𝑔𝐵2 39 74 >100
Clearly, the choice of the ideal superconducting material is highly dependent on
the application itself. YBCO might be the best choice for applications where
really high magnetic fields are required, BSCOO is the ideal candidate in
applications where the required cooling power needs to be the minimum
possible, whilst 𝑀𝑔𝐵2 could be the favourite option in cases where low cost and
high current densities are the main priorities.
2.2.2 Superconducting Components
These materials have been used in several components of generation,
transmission (i.e. power cables), distribution (i.e. transformers) and in end-use
devices such as motors. In the TeDP concept a fully superconducting network is
expected to be used. This network will consist of numerous superconducting
components, most of which will be further analysed during this research study.
Figure 18 Typical HTS Cable structure (Courtesy of Suptech.com)
33
In a power application the superconductors are consisted of multifilamentary
wires in which the superconducting filaments are embedded in a matrix of a
normal metal, an insulation system (electrical and thermal) and possibly more
layers which aim to protect the cables from magnetic flux jumps and quenching
(Larbalestier et al., 2001).
Figure 18 demonstrates the structure of a typical HTS cable. Each high power
application defines different parameter sets in regards to the critical limits of
magnetic field, temperature, and current density of these superconducting
cables. Table 3 summarises the wire performance requirements for various
industrial devices, whilst an ongoing dialogue between several science
communities continuously changes these requirements based on the
technology improvements and evolution of superconducting materials. The
three materials described in the previous section (2.2.1) are the only
superconducting materials which can satisfy these requirements.
Table 3 Industry wire performance requirements for various device applications
(Larbalestier et al., 2001)*
Application 𝑯𝒄 (T) 𝑻𝒄 (K) 𝑱𝒄 (𝑨/𝒄𝒎𝟐) 𝑰𝒄 (𝑨) Wire
Length (m)
Fault current
limiter
0.1 − 3 20 − 77 104 − 105 103 − 104 1000
Motor 4 − 5 20 − 77 105 500 1000
Generator 4 − 5 20 − 50 105 >1000 1000
Transmission
cable
< 0.2 65 − 77 104 − 105 100 per
strand
100
Transformer 0.1 − 0.5 65 − 77 105 102 − 103 1000
*According to the reference data was supplied by R.Blaugher
Generally, the commercialisation of many HTS wires have made the use of
these cables in practical applications widely possible. However, there is still
34
need for a better understanding of these materials which could lead to even
better wire performances. The reduction of AC losses, the stability
improvement, the simplification of the processing, and the improvement of the
wire robustness are only some of the areas that need further development
(Zhang et al., 2014).
The electrical machines (i.e. motors and generators) of a configuration similar to
the TeDP are expected to be fully superconducting. The motivation behind
using this type of machines is their ability to carry very high current density
without any resistance, thus enabling lighter machines. Superconducting
generators can increase the machine’s efficiency to over 99%, while
simultaneously losses can be reduced by up to 50%. These numbers are even
higher for airborne generators (Barnes, Sumption and Rhoads, 2005). A full list
of the state of the art superconducting machines will be presented in Chapter
4.1. The majority of these machines have only their rotor primarily made of
superconducting materials, whilst they consist of a conventional copper stator.
Fully superconducting machines have only been built once or twice, and little
has been published on those that have. However, in order to acquire the
required power densities for a TeDP configuration, both the rotor and the stator
will have to be superconducting for high currents, compactness and low losses.
In Chapter 4.2 a method of estimating the weight of these machines as reliably
as possible, given the current state of understanding, is described.
Other equipment that is expected to be superconducting in the TeDP
configurations under investigation are the whole transmission system (i.e.
cables), parts of the protection equipment, switching devices and possible
energy storage mechanisms. All these components will be described in detail in
the following chapter (i.e. Chapter 3.3) where the several elements of fully
superconducting networks will be explored.
Superconductors require cryogenic temperatures to operate and hence the
cooling system constitutes an important feature in this new aircraft design. The
feasibility of such a configuration highly depends on this secondary system
which will add weight and complexity to the whole architecture.
35
2.3 Cooling system
The importance of the cooling system in these novel configurations has been
highlighted in many parts of this research study. The complexity and additional
weight of this system has been considered by many as the main barrier of using
superconductors in power applications which are weight sensitive. Furthermore,
the need for excessive cooling power was the main reason why Low
Temperature Superconductors (LTS) were never broadly used.
There are two main cooling methods which have been examined as potential
cooling systems for the TeDP configurations; the use of cryo-coolers or the use
of liquid cryogen cooling.
2.3.1 Cryogenic Fluid with a Heat Sink
The main advantage of this method is the possibility of using the liquid coolant
also as a fuel for the aircraft. This will result in an almost 100% efficient cooling
system since any losses this system may have could be automatically used as
a fuel for the propulsion system. In this concept the cryogen is being loaded at
the airport in a quantity that will be enough for the flight duration in addition to
an adequate margin for safety reasons. Before each flight the tank will have to
be refilled allowing the minimum possible weight penalty for the cooling system.
In NASA studies hydrogen has been explored as a possible cryogenic fluid for
their N3-X model (Gibson et al., 2010). Hydrogen can be cleanly converted into
electrical energy through fuel cells or even by burning it in high speed turbo-
generators without any significant emissions. The disadvantage of hydrogen is
its volume; for the same fuel energy hydrogen has four times the volume of jet
fuel. Notwithstanding its volume, hydrogen has a substantially high energy unit
mass which results to only one third of the weight compared to a jet fuel of the
same energy (Felder et al., 2011a). Hence, it can be stored in liquid form at
cryogenic temperatures without adding excessive weight, providing an
adequate cooling system for the proposed designs. An example of a typical LH2
powered aircraft power system configuration is shown in the next figure:
36
Figure 19 An example of a LH2 power system TeDP configuration (Masson et al.,
2007)
In such a configuration the gas turbines run at their optimum rotational speed,
maximizing their efficiency, while coupled to high speed generators. The H2
tank provides hydrogen to the propulsion motors as well as the generators,
which in this case could be fully superconducting. Hydrogen has a boiling point
at 20.28 K which is significantly lower than the expected operating temperature
of HTS. In addition, it could potentially cool even MgB2 superconductors.
Moreover, liquid hydrogen provides an operating temperature that yields very
high current densities and as a result smaller and lighter electrical components
(Felder et al., 2011a).
Other liquid coolants such as nitrogen or methane could also be considered.
However, their boiling points (77 and 111 K respectively) complicates their use
as an exclusive mean of cooling. Methane’s boiling point is too high to be used
as the main coolant even for a HTS network, nonetheless its use should not be
excluded in the case of a double stage cryo-cooler (presented in A.2) cooling
system where it can serve as a heat sink for the cryo-coolers. On the other
hand, liquid nitrogen might not be the best alternative to hydrogen mainly
because the critical current density at 77K is typically too low to yield
competitive superconducting machines and transmission lines in terms of
weight and efficiency (Felder, Kim and Brown, 2009).
37
2.3.2 Cryo-coolers
A cryo-cooler, in simple terms, is a refrigerator that produces very low
temperatures. In the TeDP case the required temperatures are expected to be
between 20K and 65K, depending on the superconducting material that is being
used. Later, in this research study the worst case scenario in terms of
temperature (i.e. 20K) will be explored. Cryo-coolers pump the heat generated
by losses in the superconducting machines from the highest available
temperature of the device to the sink temperature where the heat is rejected.
Unfortunately, by today standards, cryo-coolers are too heavy for airborne
applications. A cryo-cooler specific mass less than 3 kg/KW of input power is
required to keep the cryo-cooler mass into accepted limits. In general, most
cryo-coolers have 3 to 5 times the desired mass (Radebaugh, 2012). The most
promising type of cryo-cooler in terms of weight seems to be the reverse
Brayton. In comparison to other active cooling configurations, turbo-Brayton
cryo-coolers produce a continuous cycle gas flow at a high flow rate. The latter
allows a constant heat transfer of high capacity from the cooling load to the heat
rejection site (Guzik and Tomsik, 2011). Figure 20 presents a survey of the
existing reverse Brayton cryo-coolers used in industrial applications. Note that
significantly lighter machines should be expected if these components are
optimised for use in an airborne application (Palmer et al., 2013).
38
Figure 20 Reverse Brayton Cryo-cooler study (Berg et al., 2015a)
Based on the study shown in the figure above an equation that links the specific
mass of the cryo-cooler (𝑚𝑐𝑟𝑦𝑜) with the input power requirement (𝑃𝑖𝑛) can be
derived:
mcryo =27.5 exp (-1.225*(log10Pin))
(2-2)
The specific mass estimated from the equation above includes the heat
exchangers, the compressors, the piping and the insulation of the cryo-cooler
and is measured in kg/kW, whilst the input power requirement in kW is based
on a Carnot efficiency of 0.3 (Berg et al., 2015a). . The amount of required
cooling power depends on the operational temperature of the superconducting
materials being used as well as the sink temperature where the heat is rejected.
The bigger the difference between these two temperatures the greater the
power required. That is the main reason why this power is larger for MgB2 than
for BSCCO based devices.
The most critical components of these components have been identified;
compressors, turbines, and heat exchangers must all show some level of
improvement over the current level of technology if goals outlined by Luongo et
39
al. (Luongo et al., 2009) are to be realized. The following figure shows the past
performance as well as the projected trendline of aerospace cryo-coolers until
the 2050 timeframe.
Figure 21 Projected development of cryo-coolers optimised for aerospace
applications (Palmer, Pagonis and Malkin, 2015)
The optimal number of the required cryo-coolers in a TeDP configuration is yet
to be decided. However, the choice of a single central cooler is already being
excluded, because of the unaccepted case of a single point failure. Each turbo-
generator should have one or even more cryo-coolers on its own, while a group
of adjacent motors in the propulsor units could probably share one cooler.
Either way, factors such as weight, safety, efficiency and cost should be taken
into consideration before the cooling system is fully decided.
In the DEAP project, models that estimate the weight of single and double stage
turbo-Brayton cryo-coolers were developed and used also in the present study.
A more detailed analysis of these models can be found in A.2 where the chosen
architecture and all the assumptions being made during the development of
these models will be pointed out.
40
2.4 Summary
Throughout the years there have been several studies focusing on the concept
of Distributed Propulsion (DP). Several versions of DP have been analysed in
this chapter which have as a target to decide the most promising configuration
in terms of weight and efficiency. The feasibility of small gas turbines DP, which
has been popular in the first DP studies, is limited mainly due to the excessive
fuel consumption associated with the small gas turbines. However, their
performance could be improved in case of significant advances in heat
exchangers technology. The configuration with distributed driven fans seems as
a more beneficial architecture. In driven fans DP concept there are three
different possible transmission systems: mechanical transmission, tip turbine
driven fans, and electrical transmission. The weight of the mechanical
transmission system and the lack of available space in the case of gas
transmission were the main barriers for the first two versions of distributed
driven fans configurations. Electrical transmission seems to be the most
promising architecture in the long term.
Turbo-electric Distributed Propulsion (TeDP) appears to be the most favourable
future disruptive technology. The weight of the electrical components is the
main drawback of such a configuration. Hence, its feasibility as a concept
depends on the availability of superconducting elements. Both NASA and
European projects such as DEAP are investigating the electric DP and the
potential superconducting nature of the proposed propulsion systems.
Based on the power requirements of the equipment in a DP configuration only
three superconducting materials could be used in this network: BSCCO, YBCO,
and 𝑀𝑔𝐵2. Although the first two have higher operational temperatures, 𝑀𝑔𝐵2
has some attractive characteristics (such as availability in wire form, lower cost
etc.) that cannot be ignored. There is however a caveat to superconducting
materials; they require cooling to cryogenic temperatures in order to perform as
superconducting.
The cooling system is clearly an important secondary system of these novel
configurations. The two options explored by the major Institutions at the
41
moment are the use of a cryogenic fluid or of mechanical cryo-coolers. Although
a cryogenic fluid could also be used as a fuel in some configurations leading to
an almost loss-free cooling system it is a technological step which requires
significant background studies and combined with the already disruptive
technology of TeDP suggested in this study might jeopardise the consolidation
of electric DP as one of the most promising future concepts. Both in DEAP
project and in the present research study the cryo-coolers’ option has been
further investigated.
42
3 Design of Autonomous Electrical Power Networks
One of the initial objectives of this research study was the design and the
simulation of the complete superconducting power network which a
Turboelectric Distributed Propulsion (TeDP) powered aircraft will implement.
However, in the early stages of this study it became clear that such a model
cannot be developed without extensive laboratory work. What makes the
modelling of such a configuration unfeasible is the fact that fully
superconducting DC networks present, at least theoretically, zero resistance
making the current sharing of the superconducting transmission lines really
difficult to be predicted. Conventional modelling strategies and tools are not
designed to simulate zero resistance systems; hence, even the design of a
steady state model for such a configuration requires additional experimental
work.
In this chapter, the main design issues and characteristics of a Superconducting
Power Network (SPN) will be described. In order to achieve this, firstly the
design process of a conventional power network will be presented. As a first
example, the design procedure of a hybrid/electric ship will be described and an
actual working example will be presented. The next step will be the description
of a SPN, emphasising its different design approach compared to the
conventional power network designs and the novel elements that such a
network include. Some of the design issues and limitations of these networks
will also be pointed out, whilst an overall synopsis of benefits and constraints of
FSNs will conclude this chapter.
3.1 Introduction to Electric Power Network Design
The design process of a SPN could become clearer if the design procedure of
an autonomous conventional power network is firstly described. This design
process is not as clear and far more complex than might have been imagined
since most of the current power networks are extensions (i.e. grid power
networks) or modifications of already existed power systems (i.e. automotive,
marine, and aerospace industries). Clearly, setting up a general network design
43
process might be misleading since the requirements of each power network
depend on the application it is designed for. However, a generic design
methodology will be described in the following section followed by a specific
working example. The ultimate goal of this section is the representation of the
power network of a TeDP type of aircraft where superconductivity will be in use.
A conventional power network with many similarities both in components and in
power rating terms is the hybrid/electric ship. Besides, modern maritime vessels
include advanced systems in many areas of interest for aircraft designers
(Bollman et al., 2015). Marine electric vehicles were occupying a $2.6 billion
market in 2013, a number that is expected to be more than double by 2023
(Harrop, 2013).
3.2 Conventional Design of Autonomous Electric Power
Networks (EPNs)
In this section a general design process of a conventional power network will be
initially described. A specific example of such a network will follow, where the
several steps of the design process will be supported by equations and
numerical examples. The main objective of this section is to point out the factors
that drive the design of a conventional power network, so that a comparison
could later be made with the superconducting version of these networks.
3.2.1 Proposed Autonomous Power Network Design Process
The following graph (Figure 22) demonstrates a design methodology of a
conventional autonomous power network. The first stage in the design process
of any power network is the analysis of the requirements. The power
requirements of each network are typically known at the very beginning even
when someone starts the design process from a blank page. Power and
operational requirements need to be specified and allocated to relevant
functional components. The power requirements are what basically size the
whole network. These determine the basic parameters of each electric network
such as the system voltage (Vs), the normal currents (Is), and the frequency (f).
Depending on the application a different selection of basic parameters can be
made even for networks of similar power requirements. Main bus bar size,
44
protection devices availability, required power converters, and overall system
losses are some of the criteria that typically determine the basic parameters
selection. More specifically, normal currents are the sizing factor for the main
bus-bar of each network. The higher these currents are the heavier the bas bar
will be. Furthermore, normal currents level also determine the fault currents
level and hence the required rating of the switchgear. Switchgear devices have
typically an interruption performance limit of around 40 kA (kilo Ampere) and
ratings above this limit should be avoided (Malkin and Pagonis, 2013).
Figure 22 Design Process diagram of a conventional power network (Malkin and
Pagonis, 2013)
Once these parameters are selected, an initial topology of the network will be
defined. This topology will include the type and number of the main power
sources (i.e. engines, generator sets, energy storage etc.), the number of
switchboard sections, the converters and transformers of the system, and
generally all the necessary equipment for the network. Depending on the
network nature and requirements different architectures could be implemented
such as bus, star, tree, ring, mesh, and/or a combination of them. The selection
between AC and DC distribution is also a challenging task which requires
45
careful consideration in the initial stages of the design process of a power
network since it dictates many of the equipment later being used.
The next stage of the design process will be an initial load analysis (i.e. steady-
state low-abstraction model) of the power network which will determine the
overall power system design. The load-flow study is basically a numerical
analysis of the electric power flow in an interconnected system. Various power
parameters of the network such as voltages, voltage angles, real and reactive
power of each bus and line of the network are determined during this load flow
analysis. These studies are important both for optimising the performance of
existing networks but also for planning future expansion of already existed
networks (Andersson, 2006). Generally the load flow problem is formulated by a
set of non-linear equations.
f (x, u, p)=0
(3-1)
Where f is a n-dimension non-linear function, x is a n-dimension vector of the
unknown component parameters (i.e. voltage magnitudes and angles in each
node), u is a vector with known parameters (i.e. machines’ voltages), and p a
vector including the network parameters (i.e. lines’ resistance and reactances).
Due to the non-linearity of the power flow analysis, this cannot be solved
analytically and hence iterative solutions are commonly implemented. Newton-
Raphson, Gauss-Seidel, and fast-decoupled-load-flow-method are just a few of
the solution methods being used to deal with the non-linear set of equations of a
network’s load flow. This analysis is of utmost importance to design the
different power system components (such as alternators, transformers,
transmission lines etc.) in order to be able to withstand any stresses they are
exposed to during their steady state operation. These stresses could be a result
of fault and short-circuit currents.
These fault conditions are typically caused accidentally through insulation
failure of the components, externally factors difficult to predict (such as lightning
strokes), or simply by faulty operations. Short circuits are the most frequent fault
46
in high power applications and depending on the location could cause stability
problems, mechanical and thermal stresses and interference with conductors.
The system needs to be protected from such faulty conditions by isolating the
faulty parts of the network as quickly as possible typically with the use of
protection devices such as circuit breakers. These devices ought to be capable
of withstanding the maximum anticipated short circuits. As an example the
calculation of a transient short circuit in a transmission line will be described
(Andersson, 2006). It is known from circuit theory that the current of a circuit is
composed of a steady state alternating current (𝑖𝑠) and a transient direct current
(𝑖𝑡):
i =𝑖𝑠 + 𝑖𝑡
(3-2)
Where,
𝑖𝑠 =√2𝑈
|𝑍|sin(𝜔𝑡 + 𝛼 − 𝜃)
(3-3)
𝑖𝑡 =√2𝑈
|𝑍|𝑠𝑖𝑛(𝜃 − 𝛼)𝑒−(𝑅
𝐿⁄ )𝑡
(3-4)
𝑍 = √𝑅2 + 𝜔2𝐿2∠(θ = tan−1𝜔𝐿
𝑅)
(3-5)
Where U is the system’s voltage, R the resistance, L the inductance, Z the
impedance, 𝜔 the frequency, t the time that the short circuit started, and θ is
the voltage angle, whilst the parameter 𝛼 controls the instant on the voltage
wave when the short circuit occurs. However, the selection of circuit breakers is
based on another short circuit value the so-called maximum momentary short
circuit current (𝑖𝑚𝑚) which corresponds on the first peak of the short circuit
47
waveform and can be as high as double the value of the symmetrical short
circuit current:
𝑖𝑚𝑚 ≤ 2√2𝑈
|𝑍|
(3-6)
Power quality is one of the main priorities while designing an electric power
network. This can be jeopardised by the distortion in the voltage and current
waveforms caused by the harmonics. The most common type of distortion is a
periodic steady-state where the distorted waveform has a Fourier series with a
fundamental frequency similar to the power system’s frequency (Ranade and
Xu, 1998). Generally, the Fourier series for a regular periodic function is given
by the following equation:
𝑓(𝑡) = 𝐶0 + ∑ 𝐶𝑛 cos(𝑛𝜔𝑡 + 𝜃𝑛)
∞
𝑛=1
(3-7)
Where 𝐶0 is the dc value of the function, 𝐶𝑛the peak value of the 𝑛𝑡ℎ harmonic
component, 𝜃𝑛 is the phase angle, whilst 𝜔 is the fundamental frequency and is
equal to 𝜔 = 2𝜋𝑓 𝑟𝑎𝑑/𝑠𝑒𝑐. An example of the synthesis of a waveform from
harmonics can be seen in the next figure:
48
Figure 23 Synthesis of a waveform from harmonics
Generally, the propagation of each harmonic is studied separately since as it
was described earlier the transmission system is typically simplified to a linear
system. A small number of harmonics is typically considered in the early design
stages. Harmonic studies are aiming on identifying the distortion levels in
voltage and current waveforms of several points of the power network and
evaluate the measures that need to be taken in order the harmonic caused
problems not to affect the power quality of the whole network. The need for a
harmonic study is typically indicated by the excessive measured distortion in
systems which include several harmonic-producing equipment (i.e.
transformers, switching devices, rotating machines etc.).
Finally, one of the main objectives of the steady-state model is the optimisation
of the protection system coordination. Protective device coordination could be
defined in simple words as the process of determining the optimal solution in
terms of current interruption in abnormal electrical condition circumstances.
Different protective zones are commonly set up in order to isolate potential
faults in small regions of the network (Glover, Sarma and Overbye, 2010).
However, the main objective of the coordination study is to minimise the outage
of any zone as much as possible.
Once the steady-state modelling is complete the key elements of the network
will then might be resized and redefined in order to optimise the network
performance during normal operation. A more comprehensive dynamic model is
49
the next step of the design process of an autonomous power network. Figure 24
summarises some of the dynamic phenomena being investigated during this
stage. C and D dynamics have already been described as parts of the steady-
state analysis.
Figure 24 Dynamic phenomena with their corresponding timescales in a power
network: A. Electro-magnetic transients, B. Synchronous machine transients, C.
Quasi steady state, and D. Steady-state phenomena (Andersson, 2006)
Transient motor starting analysis is another example of dynamic phenomenon
in an electric power network. When motors start they typically require a high
inrush of current (5-7 times their normal current) for a short period of time which
could lead to excessive voltage drop. This drop needs to be monitored and
analysed so that its effect both on the motor itself (i.e. possibility of stall) and on
other equipment to be fully understood. In most of the cases variable frequency
drives are used to overcome this initial low voltage condition (AVO, 2015).
Similar to motor starting transients transformers’ inrush currents could also
exceed the nominal current and depending on their magnitude they can affect
the power quality of the network as well as they could trip protective relays.
50
Power transformers are typically one of the most expensive components in an
electric power network and the excessive high current forces caused by these
transients could affect their life expectancy (Ebner, 2007).
Stability issues are the main concern during this dynamic stage of the design
process. The power system stability can be defined as the ability of an
autonomous power network to regain a state of operating equilibrium after being
subjected to any type of disturbance. Regaining an operating equilibrium does
not necessarily mean returning to the initial steady-state condition but to a
steady-state acceptable condition which will not result to protection actions
causing further disturbance to the system (Andersson, 2006).
The latter stage (i.e. dynamic model topology) will again redefine the system’s
requirements possibly leading to different network architectures and component
ratings. The design process is basically a constant feedback procedure on how
well a design satisfies the system requirements. Several modifications of the
initial design concept will lead to a final design that will meet the total mission
effectiveness requirements.
All the aforementioned steps can be better demonstrated with a current working
example of such a network. Hence, the design process of a hybrid/electric ship
example will be described in the following section. The selection of this network
was based on the many similarities this network share with the TeDP concept
for aerospace applications.
3.2.2 Hybrid/electric ship design process example
An example of the propulsion system of an electric ship is demonstrated in
Figure 25. This system typically consists of a number of prime movers which
provide the required electric power both for the propulsive units and for the
auxiliary loads. This electric power is transmitted to the whole power network
via generators connected to the prime movers, whilst power conversion
equipment, switchgear and the main bus-bars and transmission lines are
located between the power sources and the propulsive units (loads) to secure
the transmission of the electric power efficiently and reliably.
51
Figure 25 Electric/Hybrid Ship Propulsion System Diagram (Malkin and Pagonis,
2014)
In a hybrid ship architecture power generation modules convert fuel into
electrical power. This module typically consists either of gas turbines or diesel
engines (or even both), generators and possibly power electronics and control
modules. The majority of ships have at least two different types of power
generation sources: a main and an auxiliary one. In case that both gas turbines
and diesel engines are used as parallel main prime movers, special attention is
needed due to their different transient response. Diesel engines tend to react
faster than gas turbines in the transients and a danger of diesel overloading can
only be avoided by careful modelling and simulation (Doerry and Fireman,
2006).
52
Figure 26 Diesel-electric ship propulsion plant (marine.man.eu, 2015)
The image above demonstrates an example of a diesel-electric propulsion
plant. From the right to the left one can see the diesel engine alternators, main
switchboards, frequency converters/variable speed drives, electric propulsion
motors, gearboxes, and finally the propellers. The operation mode with the
highest expected electric load typically evaluates the type, rating and capability
of the engines. If for example the propulsion power demand of a vessel is 8MW
with a maximum consumer electric load of 2MW then the engines selection will
be driven by the following equations:
𝑃𝐵𝑝𝑟𝑜𝑝 =𝑃𝑠ℎ𝑎𝑓𝑡
𝑛𝑡𝑟𝑎𝑛𝑠=
8
0.90= 8.88 𝑀𝑊
(3-8)
𝑃𝐵𝑒𝑙𝑒𝑐 =𝑃𝑒𝑙𝑒𝑐
𝑛𝑎𝑙𝑡=
2
0.95= 2.11 𝑀𝑊
(3-9)
𝑃𝐵 = 𝑃𝐵𝑝𝑟𝑜𝑝 + 𝑃𝐵𝑒𝑙𝑒𝑐 = 10.99 𝑀𝑊 (3-10)
After the total engine brake power demand is calculated the number and type of
the diesel engines are selected based on the maintenance strategy, the mission
profile, the boundary conditions and the fact that the maximum allowed loading
of the engines should not exceed the 90%. Thus:
53
𝑃𝑇𝑜𝑡𝑎𝑙 =𝑃𝐵
𝑀𝐸𝐿=
10.99
0.9= 12.21 𝑀𝑊
(3-11)
An even number of engines is typically chosen to ensure the symmetrical
loading of the bus bars. If a number of four engines were selected in this
example then a power rating of around 3.05 MW each will be required. The
following table summarises the parameters and the values of this example:
Table 4 Diesel-electric propulsion plant main parameters
Parameters Symbol Value
Shaft Propulsion Power 𝑃𝑆ℎ𝑎𝑓𝑡 8 MW
Electric transmission efficiency 𝑛𝑡𝑟𝑎𝑛𝑠 90%
Engine brake power for transmission 𝑃𝐵𝑝𝑟𝑜𝑝 8.88 MW
Electric consumer load 𝑃𝑒𝑙𝑒𝑐 2 MW
Alternator efficiency 𝑛𝑎𝑙𝑡 95%
Engine brake power for consumer 𝑃𝐵𝑒𝑙𝑒𝑐 2.11 MW
Total engine brake power demand 𝑃𝐵 10.99 MW
Maximum engines electric loading 𝑀𝐸𝐿 90%
Total engine brake power installed 𝑃𝑇𝑜𝑡𝑎𝑙 12.21 MW
The overall electric transmission efficiency was assumed to be 90%, whilst a
relatively conservative assumption for the generator’s efficiency was made (i.e.
95%).
Power generation modules in marine applications typically produce 3 phase/60
Hz power. The standard generated voltages could be either a low voltage
450VAC system or high voltage systems typically between 4.16 and 13.8kV.
The choice between the two main voltage levels depends on the availability of
circuit breakers of sufficient rating and the total ship power demand. In many
occasions split plant operation might be chosen in order to double the total ship
power generation capability limits and increase the reliability of the vessel.
54
The power distribution system transfers the electric power to the different
network subsystems. It consists of cables, switchgear, and load monitoring and
fault protection equipment. A selection between high and low voltage buses, as
well as a choice between AC or DC electrical distribution system needs to be
made. As it was previously mentioned circuit breakers’ power rating availability
and total system generation power required favour the use of an architecture
choice over the others. Propulsion motor modules can also have an impact on
the selection of a bus voltage. Figure 27 demonstrates the recommended bus
voltage levels depending on the generation power required.
Figure 27 Bus Voltage Levels for given total required power demand (Doerry and
Fireman, 2006)
The switchboard design is mainly determined by the short circuit currents and
by the required capacity of the circuit breakers. In the previous example where
a total engine power lower than 14 MW was estimated, the 450 VAC might be
the preferable choice in regards to the system’s voltage. A rough estimation of
the anticipated short circuit levels of this example can be made by using the
following equations:
𝐼𝐺𝑠𝑐 = 𝑛 ∗ 𝑃𝐺𝑒𝑛
√3 ∗ 𝑉𝑟 ∗ 𝑥𝑑" ∗ cos 𝜑=
4 ∗ 3050
√3 ∗ 450 ∗ 0.16 ∗ 0.9= 108.69 𝑘𝐴
(3-12)
55
𝐼𝑀𝑠𝑐 = 𝑛 ∗ 6 ∗ 𝑃𝑚𝑜𝑡
√3 ∗ 𝑉𝑟 ∗ 𝑥𝑑" ∗ cos 𝜑=
2 ∗ 6 ∗ 6100
√3 ∗ 450 ∗ 0.16 ∗ 0.9= 652.19 𝑘𝐴
(3-13)
Where:
Table 5 Diesel-electric propulsion plant switchboard parameters
Parameters Symbol Units
Generator short circuit current 𝐼𝐺𝑠𝑐 kA
Number of generators/motors 𝑛 -
Rated power of the generator 𝑃𝐺𝑒𝑛 kW
Rated Voltage 𝑉𝑟 V
Sub-transient reactance 𝑥𝑑" %
Power factor cos 𝜑 -
Motor short circuit current 𝐼𝑀𝑠𝑐 kA
Rated power of the motor 𝑃𝑚𝑜𝑡 kW
In this case the circuit breaker capacity is extremely high; hence a different
voltage level shall be more appropriate for both the generators and the motors.
Typically, marine switchboards have short-circuit withstand strength of up to
150 kA (peak 330 kA) (Kongsberg, 2015). If for example generator voltages of
690 V were chosen instead, then the short-circuit levels will be reduced to 70.89
kA. A higher voltage level will be necessary for the motors where a 6.9 kV might
be the preferred choice. This would reduce the motor short circuits to
acceptable levels (i.e. around 42.5 kA). Generally, on board it is easier to deal
with lower voltages. Thus, the choice of the switchboard voltage is a trade-off
between short circuit and voltage controllability. In the previous equations sub-
transient reactance of 0.16 was assumed for both the motors and the
generators. This is a typical figure for low voltage generators, whilst a value
around 0.14 should be assumed for medium voltage machines (marine.man.eu,
2015).
56
The power conversion equipment converts electric power from the form (i.e.
voltage and frequency) of one distribution system to the form of another. It
generally consists of solid state converters and transformers. The power
conversion equipment associated with the generators and motors is typically
considered as part of the power generation system. Several methods regarding
the optimal power rating for power conversion modules are available in the
literature (Amy, 2005). In any case the total electric load demand needs to be
met in every instance with at least 95% probability.
Figure 28 presents a simplistic diagram with typical efficiencies of a hybrid-
electric ship propulsion system including the electrical machines and the
transmission system. An overall system’s efficiency of 92% is estimated, a
value which in aerospace applications will not be acceptable both for efficiency
reasons and mainly due to the derived cooling requirements. Note that any
losses derived from the generation of power where at this point neglected. A
similar diagram will be presented for the superconducting case in order to
highlight the improved efficiency these systems present.
Figure 28 Typical losses diagram of a hybrid-electric ship propulsion system
Although quite similar, the design process of an aircraft presents some main
differences. In an aircraft network the importance of lightweight components
which occupy the minimum space possible becomes more important. Minimum
weight combined with maximum efficiency is the main priority in an aerospace
application; hence superconducting networks’ attractiveness is enhanced. The
proposed TeDP concept is characterised by its superconducting nature and
57
includes some elements that have never been used before in an aircraft
system. Some of these elements will be described in the following section.
3.3 Superconducting Electric Power Network Elements
Some of the characteristics of the TeDP network have already been described
in the literature review section of this research study. A fully superconducting
network is expected to be necessary in a TeDP configuration. This network will
include several elements that are quite novel, whilst some of them have not
even been built yet. This section will present some of these novel
superconducting components, whilst the design process of a fully
superconducting network will be described in the following subchapter (i.e. 3.4).
The various innovative elements are outlined in this part of the study for
reasons of completeness. These elements combined with the exceptional
characteristics of superconducting networks are expected to notably change the
way these networks are designed.
3.3.1 Superconducting Electrical Machines
The EPN of the aircraft under investigation will consist of a number of gas
turbine alternator sets and numerous motor driven fans as propellers. Both the
generators and the motors of this configuration are expected to be
superconducting. These machines are attractive for an aerospace application
due to their significantly lower weight and volume and their extremely high
efficiency. The vast majority of existing superconducting electrical machines are
partially superconducting (i.e. superconductors used only in their rotor). These
machines are expected to provide extra weight and efficiency benefits if both
the stator and the rotor are constructed primarily by superconducting materials.
Machines with efficiency around 99.97% and two to five times better power
density than the conventional equivalents should be available in the 2035
timeframe (Brown, 2011). An extensive literature review of the already existing
superconducting machines will be presented in the next chapter (4.1), whilst a
novel method of reliably estimating the weight and volume of future fully
superconducting machines will be the main objective of Chapter 4.
58
3.3.2 Superconducting Switches
The control and switching subsystems of these networks are expected to
minimise the use of conventional mechanical switches by the use of
superconducting equivalents, which will implement local temperature and
magnetic control (Malkin and Pagonis, 2015a). Switching has to be available
not only for fault currents, but also to normal currents, when a quick
reconfiguration of the circuit is essential. These revolutionary switching devices
are expected to eliminate one of the constraints on high-current DC networks
which normally are difficult to switch due to the lack of current zeros. The main
attractive feature of these devices is their almost zero resistance which allows
them to scale-up to high operating voltages and currents without any severe
weight and conduction losses penalty. This eliminates one of the major
constraints in the design of conventional power networks which is the
switchgear capabilities and availability. Besides, superconducting switches with
fast responses have already been developed showing promising results
(Solovyov and Li, 2013).
3.3.3 Superconducting Fault Current Limiters (SFCLs)
The possibly high normal currents chosen in such a configuration will lead to
extremely high fault currents. The latter in a conventional network creates
significant problems since circuit breakers of sufficient rating both for normal
and fault operations of that extend will be difficult to become available. Instead,
in superconducting devices such as Superconducting Fault Current Limiters
(SFCLs) are expected to solve some of the fault currents design constraints.
Superconducting Fault Current Limiter (SFCLs) possibly attract the most
interest as current limiting devices in a SPN. A Fault Current Limiter is a device
that limits the prospective fault current of a network when a fault like a short
circuit occurs. The most up to date current limiters are superconducting and
they are divided into two categories: resistive or inductive. The idea of using
superconductors to hold electric power is not something new. The current
limiting behaviour of superconductors derives from their non-linear response to
current, temperature and magnetic field changes. Exceeding a limit of one of
59
these three parameters could lead these materials to lose their
superconductivity and behave as normal conductors. In a resistive FCL, which
is the most common type of limiter, when a fault current occurs the
superconductor quenches (i.e. loses its superconductivity) and the resistance
rises sharply and quickly limiting the fault current. This superconducting device
seems ideal because in the steady state has almost zero impedance whilst
when a fault current occurs this impedance rises high enough to control the
fault. After recovery of the fault, impedance goes back to zero, making the
device “invisible” again. Thus, three are the modes of a SFCL:
Normal mode
Fault-limiting mode and
Recovery period
With its relatively low cost and its capability of very sharp transition for the
superconducting to the normal state, 𝑀𝑔𝐵2 seems the most appropriate
superconductor for this type of devices (Pei et al., 2015).
In order to get a clearer view of the performance and function of a SFCL, Matlab
Simulink models and cases were developed. Figure 29 demonstrates the
Simulink model of a single phase system consisted of a 700 VAC voltage
source, a simple resistive load and a RMS Simulink block combined with a
SFCL subsystem.
Figure 29 Simulink model of a single phase system with SFCL
60
The SFCL subsystem can be seen in Figure 30; the fundamental parameters
which have been used as inputs in this resistive type SFCL can be found in
Table 6. The response time corresponds to the time needed for the SFCL to
detect and clear the fault and it is in the order of milliseconds. A triggering
current is being used as a comparison with the system’s nominal current. If the
latter is bigger than the triggering current then the maximum impedance is being
implemented to the system to control the fault currents, whereas in the reverse
case a minimum impedance of 0.01 is imposed to the system.
Table 6 Fundamental parameters of a resistive SFCL
Inputs Units Value
Response Time 𝑚𝑠 2
Minimum Impedance 𝛺 0.01
Maximum Impedance 𝛺 25
Recovery time 𝑚𝑠 20
Triggering Current 𝐴 550
The RMS value of the system’s current is being used as an input to the SFCL
subsystem, whilst the output is the result of the product between the produced
impedance and the input current. A first order filter is also used to reduce the
harmonics, whilst a controlled voltage source is used to compensate for the
voltage sag derived from the induced fault currents (Biswas, Khan and Sarker,
2013).
61
Figure 30 Simulink model of SFCL subsystem
This simple example shows the basic function of a resistive SFCL. The results
of using a SFCL in a single phase system could be seen in Figure 31. The
SFCL responds quickly enough to limit any currents higher than the triggering
current securing the stability of the whole system.
Figure 31 Single phase current waveforms in a system with and without a SFCL
A similar behaviour is expected in a three phase system. The main function of
the SFCLs in a network similar to the TeDP power network will most probably
62
be the limitation of the fault current level interruption requirements in the
minimum level possible.
3.3.4 Protection System and Converters
Generally, the zero resistance in a DC superconducting network results in a
system with minimum natural damping. Thus, faults are able to be transmitted
rapidly throughout the system reaching their peak fault currents in just hundreds
of microseconds (Ross et al., 2014). Furthermore, the possibility of quench of
the superconducting cables is another protection challenge that needs to be
addressed in this type of network.
There are two design options in regards to the protection system of a
superconducting network. The first one allows part of the system to quench so
that the derived damping (from the line resistance) to be used for reduction of
the peak of the fault currents. The second design strategy excludes the
possibility of quenching in response to a fault. Clearly, in the latter option the
protection system needs to react rapidly to isolate the fault before it reaches the
critical current value.
Furthermore, there could be three different design paths (or a combination of
two) for the protection system of such a network: to increase the fault tolerance
of the several components of the network, to limit the fault currents and their
effects using fault current limiting devices, and to mitigate the effects by using
protection devices of really fast response.
There have been studies suggesting that some converters could be used as
protection devices that can isolate the faults from the rest of the network (Baran
and Mahajan, 2007). These modern voltage-source converters can be designed
in such a way that can be more fault tolerant and be able to act fast as current-
limiting circuit breakers. However, it is still unknown if the isolation properties of
these converters will be sufficient for a superconducting network. The possibility
of quenching due to the initial high fault currents might jeopardise the reliability
of the FSN. It seems reasonable that even if such a protection configuration is
being chosen it will need to be combined with different protection technologies.
63
The use of cryo-cooled power electronics in a SPN (not only as protective
mechanisms) will be explored later in this research study (5.2.2).
If fast-acting design path is being chosen, then circuit breakers able to respond
quickly enough to the produced fault currents could potentially be used. These
circuit breakers will have to protect the converters from reverse currents
incidents and the DC link against under-voltage. The prevention of system
quench will also be one of the main functions of these devices. Studies have
shown (Fletcher et al., 2011) that solid-state circuit breakers are capable of
responding quickly enough to prevent all the fault current effects described
earlier. Their response times can reach the order of some 10s of microseconds,
significantly faster than any other circuit breakers technologies (i.e. hybrid,
electro-mechanical etc.).
Armstrong et al. (Armstrong et al., 2012) suggested a protection system with
SFCLs used in conjunction with circuit breakers. The main role of the SFCLs
will be the reduction of the magnitude of the fault currents that will consequently
lead to lower overcurrent requirements for the electrical system. Solid-state CBs
will then be used to isolate the faulted sections. Depending on the magnitude of
the derived fault currents (i.e. after the SFCLs stage) isolators could be used
instead of CBs reducing the weight and the complexity of the system. Figure 32
demonstrates this proposed configuration consisted of several different zones of
protection. The small white squares represent the CBs, whilst SFCL devices
have been placed between the generators and the converters as well as on the
DC transmission lines in order to limit the AC and DC fault currents respectively.
Figure 32 TeDP Protection System Proposed Architecture (Armstrong et al.,
2012)
64
The optimal locations of the SFCLs and CBs in this proposed protection system
are yet to be decided. Power flow studies should be conducted as soon as the
system and component inductances are known for a SPN. An overall trade
study needs to be carried out between SFCLs and CBs. The mass of the
required CBs would be decreased by using SFCLs to reduce the CBs’ fault
current interruption ratings but deciding on the optimal number, location and
ratio between these two protection devices options is a challenging task.
3.3.5 Cooling System
Furthermore, there is another caveat to the use of a superconducting network; it
requires cooling to cryogenic temperatures at all times. This adds another
heavy and complex subsystem to the already complicated power network.
There have been studies suggesting that the required cooling system is the key
technological obstacle to overcome in order the superconducting concept to
become feasible. The main options of cryogenic cooling technology have
already been described in the previous chapter; however a closer look at the
cryo-coolers’ technology will also be included in the following chapters, where
also a detailed Simulink model of a Reverse-Brayton cryo-cooler (Appendix
7.2A.2) will be used for the case studies of Chapters 5 and 6.
3.4 Superconducting Electric Power Networks Design and
Operation
In this section the main design priorities and issues of a SPN design will be
presented with an extra focus on the TeDP aircraft application. The main
differences in the design process of a conventional EPN and a superconductive
one will be clearer after this section.
3.4.1 Basic Parameters Selection
As already mentioned, the first stage of the design process of a power network
typically specifies the power requirements of the whole network. These
requirements will then dictate the basic parameters (i.e. voltage, current, and
frequency) of the power network. There are always two main design paths for
satisfying the overall power requirements of a network: either to choose a High
65
Voltage Low Current (HVLC) power system or a Low Voltage High Current
(LVHC) system.
The voltage levels in an aerospace application are typically low (less than 300V)
mainly due to Paschen’s Law (Figure 33). This law indicate that at higher
pressures the breakdown characteristics of a gap are a function of the gas
pressure and the gap length. It was found that breakdown voltage (𝑉𝐵) in Volts
could be described by the following equation (Lieberman and Lichtenberg,
2005):
𝑉𝐵 =𝛼𝑝𝑑
ln(𝑝𝑑) + 𝑏
(3-14)
Where 𝑝 is the pressure in atmospheres or bar, and 𝑑 is the airgap distance in
meters. The parameters 𝑎 and 𝑏 are constants that depend on the composition
of the gas. For air the standard atmospheric pressure is assumed 101 kPa and
the values of the constants are 𝑎 = 4.36𝑥10−7 𝑉/(𝑎𝑡𝑚 ∗ 𝑚) and 𝑏 = 12.8. Figure
33 demonstrates the Paschen’s Curve for air and two parallel copper electrodes
separated by 1 inch. According to this diagram the minimum breakdown voltage
for any product of pressure-distance is approximately 327V. However,
depending on the precise conditions a voltage level in the order of 300V is
typically being chosen as an upper limit. This practically means that an arc will
be avoided at voltage levels less than 300V at low or high altitudes. This is the
main reason why DC voltages in an aircraft are kept below this value. However,
as the power demands increase in a configuration such as the TeDP, it seems
necessary to increase the operational voltage levels so that the conductor
weight of the cables can be reduced. By increasing the network’s voltage level
the system’s current could then be decreased, for the same power
requirements, leading to transmission lines of smaller size and weight.
Nonetheless, it should be noted that possible higher voltages will require thicker
insulation. A trade-off between the conductor’s weight and the insulation added
weight is necessary in any aircraft power network. Besides, since the electric
system will be cryogenic, there is a possibility that the breakdown voltage will be
66
less sensitive to pressure and conductor distances than it is for room
temperature applications. The revised Paschen’s curve needs to be determined
so that the optimal system’s voltage and current to be decided.
Figure 33 Illustration of Paschen’s Law (Paschen, 1889)
Generally, in order to determine the optimal system voltage for the minimum
weight and maximum power capability, a complete system level study is
necessary. This system level study should include the power densities for each
component of the electrical system. Cotton et al. (Cotton, Nelms and Husband,
2007) investigated the optimal voltage selection for aerospace electrical
systems. In their study two different types of discharge were examined:
discharge around insulated wires (else known as corona discharge) and
discharge within the insulation of the wires (else known as void discharge).
Different cable options were examined such as high current (i.e. large conductor
and thin insulation), high voltage (smaller conductor, thick insulation), DC, and
AC transmission cases. The study concluded that the optimal operating point for
an aircraft power system does not necessarily imply the use of the highest
voltage possible. Cable weight and power transfer capability trade-off studies
are necessary in any proposed network architecture. Power to weight ratios of
67
non-floating DC systems proved to be the optimal solution in terms of frequency
choice (DC or AC).
In the superconducting network of the TeDP powered aircraft the selection of a
LVHC system seems inevitable, but this is actually a preferable choice for a
SPN. The main superconducting materials considered for a power application
have already been discussed in chapter 2.2.1, whilst some of the main
characteristics of superconducting cables (including critical values of current
density, temperature, and magnetic field) have been analysed in section 2.2.
Superconducting cables are characterised by their extremely high current
capability. The maximum current capability of a copper or aluminium wire is
limited around 1-4 𝐴/𝑚𝑚2, whilst superconductors with current capability of 25
𝐴/𝑚𝑚2 with potential to reach 50 𝐴/𝑚𝑚2 have already been experimentally
used in superconducting transmission lines (Xin, Han and Liao, 2006). Even
higher current densities (over 100 𝐴/𝑚𝑚2 ) have also been claimed (Masuda et
al., 2004) which is more than 100 times better than that of a conventional
copper wire. This high current capability reduces the size and cost of the
transmission lines in the TeDP type of aircraft. The zero resistance of these
cables also significantly decreases its transmission loss. The energy losses of a
superconducting cable are derived only by the AC losses which are comparable
to the magnetization loss of the superconductor itself. The transmission losses
are expected to be at least halved in the case of superconducting cable as
Figure 34 demonstrates.
68
Figure 34 Comparison between the transmission losses of a conventional and a
superconducting cable (Masuda et al., 2004)
The significantly increased electric power load could be satisfied using thin
superconducting high current cables without having to increase the system
voltages creating corona onset issues during cruise. In fact, very thin
superconducting cables might create some practical issues in regards to
making connections and mechanical support. Hence, it is the author’s belief that
in a TeDP configuration with a SPN system currents in the range of 6-30 kA will
be selected (Malkin and Pagonis, 2013).
Figure 35 summarises the anticipated losses in a SPN including the
superconducting electrical machines and the transmission losses. This graph is
used as a comparison to Figure 28 where the typical losses of a conventional
electric ship network were presented. The SPN is more than 7% more efficient
than the conventional equivalent confirming one of the most attractive
characteristics of SPNs. However, it should be noted that the cooling system
losses have not included in the figure below. Even with these losses the overall
system’s efficiency is expected to be significantly higher in the superconducting
case.
69
Figure 35 Typical losses diagram of a propulsion system using a
superconducting network
3.4.2 Current splitting
Conventional simulation modelling tools are incapable of accurately predict the
performance of a SPN. The true zero resistance of these systems results into
an unknown current sharing behaviour at the circuit nodes of a superconducting
network, especially in the case of a DC power network. In the case of an AC
superconducting system there have been studies (Malkin, 2014) showing that
the technology behind superconducting cables is a viable option for an
aerospace application. In arrangements simpler than the ones described in this
research study good current distribution has been obtained using multi-strand
MgB2 wires (Pei et al., 2012). However, it is clear that if a network similar to the
one proposed by NASA’s N3-X model is to be chosen extensive work on
investigating DC superconducting cables for aerospace electrical applications is
necessary. A system utilising these cables will obviously benefit from the almost
non-existing losses (i.e. zero resistive losses), but several issues such as
parallel current sharing have to be investigated (Malkin, 2014). In any case,
experimental work and validation processes are crucial in order these
superconducting networks to be reliably modelled and developed.
3.4.3 Electro-magnetic Forces
The high normal currents expected in these configurations might result in stray
magnetic fields and strong electro-magnetic forces. These forces exerted
between conductors can be calculated using equations (3-15) and (3-16)
which are derived from the Amperes Law and show that the forces per unit
70
length (𝐹𝑒𝑚) are proportional to the product of the conductor currents (𝐼1 ∗ 𝐼2)
divided by the distance of the wires (𝑟).
𝐹𝑒𝑚 = 2 ∗ 𝑘𝑎 ∗𝐼1 ∗ 𝐼2
𝑟
(3-15)
Where, 𝑘𝑎 =𝜇0
4∗𝜋= 10−7 𝑁 𝐴−2
(3-16)
Hence, the system needs to be designed in such a way that it will cater for
these strong forces. The interference between the several superconducting
components is also unknown since there has never been an application
including so many novel superconducting devices all together. The nature of
this application (i.e. aerospace application) creates even more unknowns
derived from the uncertainty of the superconducting components’ performance
in altitude.
3.5 Summary
The design process of an autonomous power network is a relatively unknown
procedure since most of the current networks are extensions or modifications of
already well-established networks. This uncertainty becomes even more
profound in the case of a superconducting power network due to its novelty and
revolutionary design aspects it presents. Particularly, in the case of a
demanding power network such as the one required for the TeDP type of
aircraft the existing basic parameters standards seem insufficient to address the
power quality of such a complicated network. Due to the low TRL of this
concept, optimised standards for the TeDP system have not been set up
making the sensitivity studies of the proposed configuration a challenging task.
This is also enhanced by the superconducting nature of the electrical power
network.
71
There are undeniable benefits of using a FSN mainly derived from the improved
efficiency, power density, as well as the flexibility these networks could offer.
Nevertheless, there are still many aspects of these networks that have not been
fully understood by the aerospace industry. The full potentials of a FSN can be
obtained only if a different approach on the design process of future aircraft is
followed. Several design constraints such as cable size, switching and fault
current limiting capabilities are eliminated in the case of a FSN. On the other
hand, several design issues of superconducting networks need extra attention
since various novel components might be used concurrently for the first time in
such a sensitive network
Since superconductivity appears to be as one of the main enablers of promising
concepts such as the TeDP approach it is essential that further research studies
and significant funding resources to be dedicated on the experimental analysis
of these networks.
72
4 Superconducting Electrical Machines
One of the main components in a fully Superconducting Power Network (SPN)
is clearly the electrical machines. There have been many studies and
experimental work in regards to the construction of such machines. However,
the vast majority of the superconducting machines which have been built have
only their rotor superconducting while the stator remains conventional. In this
chapter, initially the most important examples of tested superconducting
machines will be summarised. After that, a novel method of estimating the
weight of fully superconducting machines will be presented, while a sensitivity
study focusing on the main parameters of these models will also be carried out.
Finally, the model limitations as well as its validation references will conclude
this chapter.
4.1 Status and State of the Art
The idea of building a superconducting electrical machine has been around
since the discovery of the superconductivity phenomenon (1911) when many
researchers started considering the possibility of constructing a
Superconducting Machine (SM) predicting the possible benefits that such a
configuration could offer. Until recently, it was not possible to build and test a
fully superconducting machine and the majority of the tested machines
compromise a “superconducting” rotor (with HTS or LTS materials) and a
conventional stator. In this subchapter a brief description of the machines that
have been built and tested throughout the years will be presented.
4.1.1 Superconducting Synchronous Machines
The first studies go back in the 70’s, where the development of high-field
superconductors triggered the interest for using such conductors in the
electrical rotating machines.
In 1974 the Westinghouse Electric Corporation got involved in the design,
development and test of a 10MVA, 12000 RPM AC generator with a
superconducting field winding (Blaugher, Parker and McCabria, 1977). A
73
predicted reduction in weight and volume of these machines was the motivation
behind this work. The prototype superconducting rotor of this machine was
operating in the temperature of 4.2K (LTS) while a conventional construction of
stator was chosen. Although, there were some benefits using LTS in these
machines it was not until the discovery of HTS that extended studies begun.
It was then almost twenty years later (1993) when American Superconductor
Corporation decided to design and construct two synchronous motors using
HTS field coils (Joshi et al., 1995). Both motors had a silent pole field structure
excited by HTS (i.e. BSSCO) coils which remained superconducting during all
the operation modes. The motors produced 1.5kW and 3.5kW power, operating
at 3600 (two poles) and 1800 (four poles) RPM respectively. The potential
energy savings was the main motivation behind this study since initial analysis
has showed possible increase in the overall efficiency of the motors in the range
of 1.3%. This research was really important because it proved the feasibility of
the whole concept and it formed the basis for further studies to come.
In 1997, a research group from Tampere University (Finland) used the Bi-
2223/Ag coils manufactured by American Superconductor Corp. to test a 4-pole
synchronous machine in different operating temperatures (4.2K to 77K)
(Eriksson et al., 1997). They ended up constructing a 1.5kW machine at 1500
RPM. A different approach in the design of the machine was chosen. More
specifically, the so-called inside out concept was followed where the excitation
is happening on the stator side whilst the rotor side armature operates at room
temperature. This choice was proven successful mainly for moderate power
levels and operating temperatures around 20K (77K was proved to be too
inefficient for the wires of that era).
Around the same period the U.S. Navy started to investigate the possibility of
using superconducting synchronous motors for ship propulsion (Gamble et al.,
2002). The conceptual design of a 25MW motor can be seen in Figure 36. The
study concluded that significant efficiency and noise benefits can be achieved
with such a design. Many more studies for marine superconducting propulsion
motors followed both from U.S. Navy and from other organisations. The former,
74
in 2000, developed a 5MW, 230 RPM HTS AC synchronous motor for electric
ship propulsion (Eckels and Snitchler, 2008). The reduced size and weight of
this machine allowed a more flexible design, simple installation and
maintenance, while the performance both for the steady-state and the transient
modes was significantly improved. This machine was the baseline for even
larger machines. U.S. Navy finally tested a 36.5 MW HTS propulsion motor in
2006 operating at 120 RPM demonstrating an overall efficiency of 97.3% and
concluded that the use of superconducting machines can offer architecture
benefits in both existing and new ship designs (Gamble, Snitchler and
MacDonald, 2011).
Figure 36 25MW 120 RPM superconducting synchronous motor U.S Navy
conceptual design (Gamble et al., 2002)
A recent example of marine superconducting propulsion motor study comes
from the University of Shahrood in 2014 where a design process of a HTS rim-
driven synchronous motor for marine propulsion has been developed
(Hassannia and Darabi, 2014). A 2.5MW, 220 RPM machine has been studied
and ways to reduce the axial length of these machines have been proposed.
Siemens also showed some interest in the development of superconducting
synchronous machines (Figure 37). In 1999, a HTS four-pole 400kW machine
was built at a rated speed of 1500 RPM and performed on an overall efficiency
of 96.8% (including the refrigerator) (Gieras, 2009). Two more HTS machines
were tested in the following years (2002-2010), a 4MW / 3600 RPM and a 4MW
75
/120 RPM respectively (Wolfgang, Grundmann and Frauenhofer, 2012).
Compared to a conventional generator of the same rating the former machine’s
efficiency raised from 96.1 to 98.4%. In general, both studies concluded that
HTS technology can be the solution for a more sustainable future and especially
in cases such as the marine applications a complete system re-design could
capitalise the numerous benefits that superconducting machines can offer. The
latter statement also stands for the TeDP concept, where a novel systems level
approach, and not a simple component technology improvement, is necessary.
Figure 37 Siemens HTS Synchronous Machine Test Bed (image courtesy of
Siemens)
The East Institutes showed an interest in superconducting machines both
synchronous and homopolar DC (4.1.2). For the former type, Korea Electro-
technology Research Institute (KERI) was the first one to develop a 100hp, 4
poles, and 1800 RPM synchronous motor with HTS (Bi-2223 tape) field coils in
2002 (Kwon et al., 2005). In 2005 Japanese Industry Academia Group built an
axial gap-type brushless HTS synchronous motor using Gd-Ba-Cu-O bulk
magnets for the rotor. The construction of a 3.1 kW and 720 RPM SM was
feasible (Figure 38), while possible improvements using successive pulsed
magnetization were claimed (Matsuzaki et al., 2005).
76
Figure 38 HTS Motor using Gd-Ba-Cu-O bulk magnets schematic illustration
(Matsuzaki et al., 2005)
One year later an eight-organization joint team leading by Shikawajima-Harima
Heavy Industries Co. and Sumitomo Electric Industries developed the first liquid
nitrogen cooled fully superconducting motor for ship propulsion as it can be
seen in Figure 39 (Takeda, Oota and Togawa, 2006). The rated output power
was 12.5 kW running at the speed of 100 RPM. The study revealed some very
optimistic results ending up with a machine two times lighter than the
conventional equivalent, with high efficiency, and no noise or flux leakage.
Figure 39 The first fully superconducting motor (Takeda, Oota and Togawa, 2006)
Finally, in 2007 a joint Research and Development Group funded by New
Energy and Industrial Technology Development Organization developed a 15
77
kW, 360 RPM HTS superconducting motor (Iwakuma et al., 2007). YBCO
superconducting tape was used for every field coil, whilst a copper winding with
an iron core was used for the stator. The final motor was tested as ship
propulsion system where a quite stable operation was verified.
YBCO bulk material was also incorporated into the rotor of a superconducting
reluctance motor which was successfully built and designed by a German
Institution (i.e. Oswald Elektromotoren) in 2004 (Oswald et al., 2005). A 200
kW, 3000 RPM HTS reluctance motor was finally constructed and tested
(Figure 40) with results that proved the potential use of such a machine in future
applications where high power density, small size and high dynamics are
required. However, improvements in the bulk YBCO material should be
expected and more extensive studies are anticipated.
Figure 40 200 kW HTS Reluctance Motor (Oswald et al., 2005)
Another important study that compared in terms of efficiency and size the
performance of a 1000hp HTS motor with a similarly rated conventional
machine was carried out by a U.S Department of Energy funded program
(Dombrovski et al., 2005). Four coils wounded with a multi-filament BSCCO
tape were used for the field winding, whilst the armature winding was designed
for room temperature operation. The reduction of the core-end losses and the
interaction between HTS motors and the power converters were some of the
issues that have been pointed out as challenges for the future designs.
Finally, in UK a 100 kW HTS synchronous motor was fully constructed in 2004
in University of Southampton and a systematic test program to characterise its
78
performance was also developed (Wen et al., 2009). The machine operated at
liquid nitrogen temperature (77K) having a superconducting winding consisting
of ten Bi2223 pancake coils and a 3-phase conventional stator. The schematic
layout of this machine can be found in Figure 41. It was found that the critical
current of the rotor coil significantly increases as the temperature decreases.
The field of the stator winding on the other hand does not affect the critical
current in a similar extend.
Figure 41 Layout University of Southampton’s 100 kW HTS machine (Wen et al.,
2009)
4.1.2 Homopolar DC Superconducting Machines
There were several companies such as General Atomics who claimed that
homopolar DC SMs can be superior to the AC equivalents (Gieras, 2009).
Benefits such as less noise, smaller size, better efficiency, less cost, and
simpler architecture and control have been claimed. In 1995, they successfully
demonstrate an electric motor (Figure 42) which uses superconducting field
windings constructed with BSCCO-2223. The motor was tested for two different
operating temperatures: in liquid helium temperature (i.e. 4.2 K) and in liquid
neon temperature (i.e. 28 K). Eventually a 125 and a 91 kW SM were produced
running at the speed of 11,700 RPM (Waltman and Superczynski, 1995).
79
Figure 42 HTS DC Homopolar Motor (image courtesy of General Atomics).
The baseline for the aforementioned DC Homopolar SM was a 3.7 MW
subscale motor at 500 RPM which utilized two NbTi superconducting coils that
could be easily transitioned to HTS materials (Thome et al., 2002).
4.1.3 Superconducting Induction Machines
Significant efforts and funding from the East have been dedicated in the
investigation of superconducting induction motors and generators. More
specifically, a superconducting induction generator (SIG) with HTS bulk magnet
has first been presented from Seoul National University in 2000 (Kim and Hahn,
2000). The machine consisted of two rotors, an outer one which was made of
copper and the inner one which was constructed of HTS bulk magnets. The
study concluded that the construction of a 2KVA SIG is feasible with a stable
electrical and mechanical performance.
In 2003, the Ministry of Science and Technology of the Republic of Korea
funded the study of a HTS induction motor (Sim et al., 2004). The motivation
behind this research study was the possible efficiency benefits that an induction
motor with HTS tapes as rotor bars could offer. A 0.75kW HTS induction motor
was finally constructed with HTS tapes (BSSCO-2223) used as the short bars
and rings and a comparison with a conventional motor of similar rating was
carried out. In this configuration, the superconducting bars should quench
during the starting phase to provide high current, while they recover from
quench during the normal conditions in order to improve the overall efficiency.
They concluded that the superconducting machine performed better than the
80
conventional equivalent motor with almost double starting torque and better
efficiency during normal mode.
Finally, in 2006 a research group from Japan fabricated and tested a 1.5kW 3-
phase HTS induction motor where both the rotor bars and the end rings were
made of Bi-2223/Ag multifilament tapes (Nakamura et al., 2006). A schematic
diagram of the test area is being presented in Figure 43. It was shown, both
theoretically and experimentally, that the HTS induction motor requires
minimum voltage for starting and could produce higher starting and accelerating
torque compared to the conventional motor.
Figure 43 Schematic diagram of the test system of a fabricated HTS induction
motor installed in a metal cryostat (Nakamura et al., 2006)
4.1.4 Programmable Superconducting AC Machine (PSAM) Project
PSAM project findings were used to develop a baseline machine prototype
during the DEAP project. This programme was a partnership between Rolls
Royce plc, Magnifye Limited, Cambridge University and EADS Innovations
Group (later renamed to Airbus Group Innovations). The main objective of this
technology demonstration project was initially the testing and integration of a
doubly-superconducting AC machine so that the extra potential benefits over
singly superconducting machines to be explored. The expected benefits in
weight and efficiency of these machines could make them an attractive option
for the aerospace sector. The full integration of the superconducting rotor and
81
stator in the same machine was not fully accomplished but the revised scope of
the project which describes the outline design, some basic assumptions and
some first mass estimation of these machines was successfully presented (Berg
and Dodds, 2013). The machine architecture can be seen in Figure 44, where
an environscreen is also included. The impact of this screen on the total weight
of the machine will be calculated in one of the following subchapters (4.3.1).
Figure 44 PSAM Machine Arrangement (Berg and Dodds, 2013)
This study is probably the only one considering the use of Magnesium Diboride
(MgB2) wires for the stator of these machines. Unlike other HTS, MgB2 is a low
cost superconductor available in wire form. This enables a MgB2 coil to be
constructed where the wires are transposed to enable AC operation.
The project concluded that significant improvements in the power and torque
densities can be achieved by using doubly-superconducting machines.
However, the complexity of the whole system required (i.e. need for cryo-
cooling system) in the aerospace applications dictates that the use of such
machines could be more beneficial in the high power applications where the
gains are major.
4.1.5 Summary
Singly superconducting HTS machines can be considered as mature
technology (TRL 6) for certain applications such as the marine industry for
example where a significant number of successful prototypes are already in
existence. On the other hand, doubly superconducting machines are starting to
82
attract the interest of aircraft manufacturers since they seem ideal for innovative
future concepts similar to the TeDP approach. The discovery of MgB2
superconductor is considered by some research groups the enabler to design
fully superconducting machines (4.1.4). Building such a machine it is expected
to bring extra benefits regarding their efficiency and their overall weight. Singly
superconducting machines have already been proved more efficient with higher
energy densities than the conventional equivalents. The following table
summarises the already tested superconducting machines with available
efficiency and overall weight data.
83
Table 7 List of singly superconducting electrical machines
Type Output Power (kW)
Rated Speed (rpm)
Efficiency (%)
Weight (kg)
Reference
1 Synchronous Motor
(marine)
15 360 - 900 (Iwakuma et al., 2007)
2 Synchronous Motor
(marine)
5000 230 96 23000 (Eckels and Snitchler,
2008)
3 Synchronous Motor
(marine)
36500 120 97.3 75000 (Gamble et al., 2011)
4 Synchronous Generator
4000 3600 98.7 7000 (Wolfgang et al., 2012)
5 Synchronous Generator
4000 120 96.2 36000 (Wolfgang et al., 2012)
6 First High-speed
Generator*
10000 12000 - 426 (Blaugher et al., 1977)
7 Homopolar DC motor (marine)
3700 500 ~97 11400 (Thome et al., 2002)
8 Synchronous Motor
(marine)
25000 120 97.5 70000 (Gamble et al., 2002)
9 Synchronous motor
746 1800 97.1 6000 (Dombrovski et al., 2005)
*Note that this machine has not been included in the following graph due to its oldness and its
use of LTS materials
Summarising the information of Table 7 it was possible to derive some
interesting results concerning the size of these machines. It is reasonable to
use torque (T) as the parameter that sizes these machines. The maximum
electromagnetic power at the air gap can be converted into mechanical power
(P) as:
𝑃 = 𝑇 ∗ 𝜔𝑚 (4-1)
84
Where 𝜔𝑚 is the speed in rad per second (N: number of rotations):
𝜔𝑚 =2 ∗ 𝜋
60∗ 𝑁
(4-2)
Figure 45 presents the results of this study. By using these eight machines, it
was possible to produce an equation which links the total weight of a singly
superconducting machine with its torque.
Figure 45 Weight vs Torque of singly superconducting machines
Although one might have expected a more linear relationship between the
torque and the weight of these machines this apparently is not the case for the
singly superconducting machines. The exact reasons for this unexpected
trendline are not yet clear to the author and further research on this subject
might be necessary. The derived equation (4-3) could only be used as an early
stage indication for the weight of singly SMs. Its relatively low 𝑅2 value does not
1
2
3
4
5
7
8
9
y = 75.541x0.4739 R² = 0.9683
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
100000
1 10 100 1000 10000 100000 1000000 10000000
We
igh
t (k
g)
Torque (N*m)
85
allow the use of this equation as a reference to the methods described in the
following chapters (4.2).
𝑇𝑜𝑡𝑎𝑙 𝑊𝑒𝑖𝑔ℎ𝑡 = 77.146 ∗ 𝑇𝑜𝑟𝑞𝑢𝑒0.4701
(4-3)
4.2 Weight Estimation of Fully Superconducting Machines
One of the most important issues of the TeDP concept is the weight of the
electrical system. The fact that fully superconducting machines have only been
built once or twice, and little has been published on those that have, makes the
weight estimation of this system even more challenging. There has not been a
study describing a reliable method of calculating the weight of these machines
and in the majority of the TeDP related journal and conference papers predicted
values are included without a clear description of the methodology behind them.
In this chapter a novel method for calculating the weight of fully
superconducting machines is described and corresponding Simulink models are
presented.
4.2.1 Torque per unit of rotor volume (TRV) method
Torque per unit of rotor volume (TRV) is what generally characterises the size
of electrical machines. This effectively depends on the product of the electric
loading (A) and the magnetic loading (B). Both these values are limited by the
properties of the materials being used as well as the temperature rise and
cooling system capability (Hendershot and Miller, 2010).
Electric loading is defined as the linear current density around the airgap
circumference and is given by the following equation:
𝐴 =𝑡𝑜𝑡𝑎𝑙 𝑎𝑚𝑝𝑒𝑟𝑒 𝑐𝑜𝑛𝑑𝑢𝑐𝑡𝑜𝑟𝑠
𝑎𝑖𝑟𝑔𝑎𝑝 𝑐𝑖𝑟𝑐𝑢𝑚𝑓𝑒𝑟𝑒𝑛𝑐𝑒=
2𝑚𝑇𝑝ℎ𝐼
𝜋𝐷 𝐴/𝑚
(4-4)
86
, where m is the number of phases, 𝑇𝑝ℎ is the number of turns in series per
phase, I is the RMS phase current, and D is the diameter of the airgap. In these
first calculations there is no distinction between the rotor outer diameter and the
stator inner diameter assuming that the airgap is too small compared to the
rotor diameter.
Magnetic loading on the other hand represents the average flux density over the
rotor surface. In AC motors this is distributed sinusoidally and is derived from
the following equation:
𝐵 = 𝛷 ∗2𝑝
𝜋𝐷𝐿𝑠𝑡𝑘 𝑇
(4-5)
, where Φ is the fundamental flux/pole, p is the number of pole pairs and 𝐿𝑠𝑡𝑘 is
the stack length. Generally, in a slotted stator of a conventional electrical
machine the peak flux density is limited by the saturation losses which can be
excessive for density values above 1.6T. This is not the case for the
superconducting machines where this value is anticipated to be at least twice as
much. Note that since it is assumed that the flux is sine-distributed, its average
value will be given as:
𝐵𝑎𝑣𝑔 = 𝐵𝑚𝑎𝑥 ∗2
𝜋 𝑇
(4-6)
Next step will be to use the standard equation that gives the generated electro-
magnetic force (emf) per phase:
𝐸 =2𝜋
√2∗ 𝑘𝑤𝑇𝑝ℎ𝛷𝑓 =
𝜋2
√2∗
𝑘𝑤𝑇𝑝ℎ𝐵𝑎𝑣𝑔𝐷𝐿𝑠𝑡𝑘𝑓
𝑝 𝑉
(4-7)
, where f is the fundamental frequency and 𝑘𝑤 is the fundamental harmonic
winding factor.
87
The maximum electromagnetic power at the airgap is given by:
𝑃 = 𝑚𝐸𝐼
(4-8)
However, power could also be derived by equation (4-1) and rotational speed
by equation (4-2). The rotor volume on the other hand could easily be
calculated as:
𝑣𝑟 = 𝜋 ∗ 𝐷2 ∗ 𝐿𝑠𝑡𝑘/4 (4-9)
Combining all the above equations (4-1)-(4-9) it is possible to come up with the
final equation that sizes the electrical machines:
𝑇𝑅𝑉 =𝑇
𝑣𝑟=
𝜋
√2𝑘𝑤1𝐴𝐵 𝑁𝑚/𝑚3
(4-10)
4.2.2 Relationship between rotor and stator dimensions
Generally, the torque of a load is commonly given as a requirement before
designing a machine. After that, an electrical machine capable of driving this
load needs to be designed. Equation (4-10) could be then used to calculate the
volume of the rotor. The next step should be the estimation of the active weight
of the machine. In order to do the latter a relationship between the rotor and the
stator dimensions need to be found. There have been some studies (Miller,
1989) suggesting that for a rough estimation of the stator dimensions a “split
ratio” (rotor diameter/stator diameter) parameter can be used. The equation
normally being used is the following:
𝑆𝑡𝑎𝑡𝑜𝑟 𝑣𝑜𝑙𝑢𝑚𝑒 =𝑅𝑜𝑡𝑜𝑟 𝑣𝑜𝑙𝑢𝑚𝑒
𝑆𝑝𝑙𝑖𝑡 𝑟𝑎𝑡𝑖𝑜2
(4-11)
88
However, the assumed value of this ratio varies with the type of the machine
and the architecture (interior-rotor or exterior-rotor) being chosen. Especially for
the superconducting machines further assumptions are necessary jeopardising
the reliability of this method. Hence, a more reliable technique to determine the
relationship between the stator and rotor dimensions was essential.
The method which has been used in this project is based on a literature survey
of actual real conventional electrical machines that gave us a reliable
relationship between the rotor and stator diameter. More specifically, 13
different conventional machines were included in this study (Appendix A.1),
while the dimensions of three of the superconducting machines which have
been presented in 4.1, as well as the PSAM conceptual baseline design (4.1.4),
were added to the graph (Figure 46) in order to validate the reliability of the
derived relationship for the superconducting case.
Figure 46 Rotor vs. stator dimensions relationship graph
As it can be seen, two relationships were derived from the graph above: one
using the power trendline in the excel toolbox and another one using a linear
relationship. These two relationships were used in the following subchapters to
estimate the weight of the fully superconducting machines by two different
methods. To summarize the aforementioned equations:
y = 2.8066x0.8874 R² = 0.9887
y = 1.144x + 83.03 R² = 0.9891
100
1000
10000
50 500 5000
Stat
or
(mm
)
Rotor (mm)
conventional machines
PSAM Baseline Design
746kW HTS Motor(Rockwell Automation)
1.5kW HTS synchronousmotor (Finland)
10MW LTS Generator(Westinghouse)
Power (conventionalmachines)
89
𝑆𝑡𝑎𝑡𝑜𝑟 𝑑𝑖𝑎𝑚𝑒𝑡𝑒𝑟 = 2.8066 ∗ 𝑅𝑜𝑡𝑜𝑟 𝑑𝑖𝑎𝑚𝑒𝑡𝑒𝑟0.8874 (4-12)
, and
𝑆𝑡𝑎𝑡𝑜𝑟 𝑑𝑖𝑎𝑚𝑒𝑡𝑒𝑟 = (1.144 ∗ 𝑅𝑜𝑡𝑜𝑟 𝑑𝑖𝑎𝑚𝑒𝑡𝑒𝑟) + 83.03 (4-13)
The superconducting machines being included in this study validate that a
similar trend will most probably be followed in the design and construction of
superconducting machines.
4.2.3 Basic Assumptions
In order to proceed with the weight estimation of the fully superconducting
machines some basic assumptions need to be made. All the following
assumptions were validated by literature studies and expert’s opinions, whilst
sensitivity studies will also follow, testing the impact of these assumptions on
the overall weight estimation of the machines.
Magnetic loading B
During the first simulations a maximum air gap flux density of 3T was assumed.
This value was derived from the PSAM project outcomes and it could be
considered a conservative estimate itself especially for the 2035 timeframe.
Even recent experimental studies (Rada et al., 2015) have suggested magnetic
field capabilities of around 4T. It is important to point out that in superconducting
machines the magnetic loading has a slightly different meaning than in a
conventional machine. In the latter, it indicates the working level of the flux
density in the air gap measuring the flux density in the iron teeth. On the other
hand, in the superconducting machines this magnetic loading is not constant
across the air gap and it is just a reference of the peak value of flux at the
armature conductors. For this reason equation (4-6) is being used in the
calculation of the overall weight of the machines. The peak value of magnetic
loading in this case does not only affect the power output of the machines but is
90
also an indication of the eddy currents loss in the armature conductors (Bumby,
1983).
Electric Loading A
The assumed electric loading value was 400 𝑘𝐴𝑚⁄ also derived from the PSAM
experimental results. This value could also be considered as conservative
especially if you compare it with higher estimates that were assumed in studies
even back in the 70’s (Miller and Hughes, 1977) as well as with more recent
studies that claimed values even in the range of 700kA/m and more (Tixador
and Daffix, 1997).
Winding factor 𝑘𝑤
The fundamental harmonic winding factor 𝑘𝑤 could be described as a reduction
factor of the generated RMS voltage in 3-phase AC generator. In most
conventional machines this factor varies from 0.85 to 0.95 (Skaar, Krovel and
Nilssen, 2006). There is no reason to believe that this will be any different in a
superconducting machine and hence a value of 0.9 was initially assumed.
Machine design and materials assumptions
Figure 47 General view of the DEAP superconducting electrical machine
(Courtesy of the DEAP project)
Figure 47 demonstrates a general overview of the machine’s arrangement used
in this project. It is basically a more detailed and advanced version of the one
91
being used in the PSAM project (Figure 44). In this design, the stator is
constructed primarily of epoxy concrete with a density of 2400 𝑘𝑔/𝑚3, while the
rotor is constructed primarily of steel (density of 7859 𝑘𝑔/𝑚3) with a small
proportion of HTS materials. As shown in Figure 44, the active rotor also has a
hollow space of the same diameter as the rotor shaft. The ball bearings are
assumed to be constructed of ceramic with a density of 4700 𝑘𝑔/𝑚3and are
70% solid by exterior volume. Vacuum chamber and vacuum chamber end
flange are assumed to be made primarily of aluminium alloy, the use of which
appears feasible at cryogenic temperatures (Senkov, Bhat and Senkova, 2004).
All the aforementioned material assumptions combined with the presence of the
hollow space inside the rotor led to an assumed active density value of
3000 𝑘𝑔/𝑚3. By active density, we consider the overall power density of the
active parts of the machine (i.e. rotor, stator and possibly environmental
screen). Finally, an extended analysis about the environmental screen of this
machine and its effect on the overall machine weight will follow in the sub-
chapter 4.3.1.
4.2.4 Models Description
In this project two different methods estimating the weight of the
superconducting machines were proposed. Both versions were based on the
TRV concept, but whilst in the first approach the outcome of the TRV equation
(4-10) was simply given the rotor dimensions, in the second version which shall
be considered as more optimistic a further assumption was made. It is believed
that for fully superconducting machines the TRV equation is linked with the
mean stator winding diameter and not the outer rotor diameter (Berg and
Dodds, 2013). Table 8 summarises the inputs and outputs of the two models
which were developed in MATLAB Simulink to estimate the overall weight of the
fully superconducting machines.
92
Table 8 Inputs/Outputs of the Simulink models for the weight estimation of fully
superconducting machines
Inputs Units Outputs Units
Output Power 𝑊 Torque 𝑁 𝑚
Rotational Speed 𝑟𝑝𝑚 Phase Current 𝐴
Electric Loading 𝐴 𝑚−1 Phase Voltage 𝑉
Peak Magnetic Flux 𝑇 Active Length 𝑚
Winding Factor − Active Diameter 𝑚
Pair of Poles − Active Volume 𝑚3
Efficiency − Active Weight 𝑘𝑔
Mean Stator Factor − Frequency 𝐻𝑧
Number of Turns − Thermal Load 𝑊
Power Factor − Power Density 𝑊 𝑘𝑔−1
Active Density 𝑘𝑔 𝑚−3 Torque Density 𝑁 𝑚 𝑘𝑔−1
Cryostat Weight Factor − Cryostat Added Weight 𝑘𝑔
Length/Diameter Ratio − Total Weight 𝑘𝑔
First Version (TRV Original)
In this first version, from equation (4-10) the rotor volume can be estimated. By
assuming an aspect ratio of one (rotor length L=rotor diameter D) and by using
equation (4-12) the stator diameter is calculated. Choosing an L/D ratio of unity
is a common technique in the initial sizing estimates of electrical machines
(Hendershot and Miller, 2010). The active dimensions of the machine can now
be estimated and by assuming an active power density of 3000 𝑘𝑔/𝑚3 the
active weight can also be calculated. The following figure presents the Simulink
model which includes all the aforementioned assumptions and equations:
93
Figure 48 Simulink model (first version) for the weight estimation of fully
superconducting electrical machines
In this model the power and rotational speed are used as inputs. In most
applications it is common to know the power and speed requirements of the
system; hence the required torque can be easily calculated (4-1). Apart from
the active weight of the machines, further outputs of the model are the following:
phase voltage, frequency, thermal load, power and torque density, and overall
weight of the machine. The phase voltage of this machine can be found by
using the well-known relationship:
𝑉𝑝ℎ =𝑃
√3 ∗ 𝐼𝑝ℎ ∗ cos 𝜑
(4-14)
94
Where 𝐼𝑝ℎ is the RMS phase current, and cos 𝜑 is the internal power factor
which in our case is assumed to be equal to one. The RMS phase current can
be derived from equation (4-4), assuming a value for the number of turns per
coil (i.e. in our models assumed as 70). The machines’ frequency is being
based on the number of pair poles (assumed as 1 in the first estimates), while
for the thermal load calculation an efficiency of 99.9% is assumed. Although this
assumed efficiency might seem relatively high, equivalent NASA studies
(Brown, 2011; Felder et al., 2011a) assumed efficiencies up to 99.97%, making
this project’s assumptions relatively pessimistic.
Finally, to calculate the total weight of the machine an extra cryostat weight
percentage is considered. The latter was chosen based on an expert’s opinion
(i.e. Steven Harrison-formerly of Scientific Magnetics) who suggested that a
value between 10-50% of the actual machine should be added. However, since
we are looking at aerospace applications a value closer to the lower limit of this
range (10-20%) seems reasonable. A 15% cryostat added weight was chosen
for Chapter 4 calculations, whilst a more conservative value of 30% was
selected for the following chapters (i.e. Chapters 5, 6).
Second version (TRV optimistic)
In this version, equation (4-10) is used to calculate the mean stator diameter
instead of the outer rotor diameter. A way to calculate the outer stator diameter
was then needed and this became feasible by using the linear trendline
equation (4-13). An additional assumption was necessary to calculate these
dimensions. More specifically, a mean stator factor was added with an initially
assumed value of 0.66. This factor basically expresses the mean stator radius
relative to the outer rotor and stator radius. A value close to zero means that the
mean stator radius is exactly the same as the outer rotor one, something that
will basically lead to similar results with the first version of the model (the TRV
equation in this case will give the rotor dimensions). Values close to one will
result in very optimistic estimates regarding the overall weight of the machines.
95
Figure 49 Simulink model (second version) for the weight estimation of fully
superconducting electrical machines
The figure above shows the Simulink model of the second version, where the
main difference with Figure 48 is the aforementioned way of calculating the
active weight of these machines. Figure 50 demonstrates this new subsystem.
96
Figure 50 Mean stator to outer stator subsystem
The rest of the model, as well as all the assumptions being made in the first
version of the model remain the same. A sensitivity study that will show the
effect of these assumptions on the overall weight of the machine will follow on
the next subchapter.
4.3 Sensitivity Study
As it is already mentioned in the previous chapters, in order to develop a model
to estimate the weight of the fully superconducting machines some basic
assumptions had to be made. The justification for these assumptions was
described in chapter 4.2.3. In this subchapter however, the effects of the most
important assumptions being made on the overall estimation of the weight will
be explored for the two different versions of the model. A sensible range for
every assumption will be investigated whilst the rest of the parameters will keep
their initial assumed value during the progress (Table 9).
97
Table 9 Initial assumed values for the model’s inputs
Parameter Units Value
Maximum Magnetic Loading 𝑇 3
Electric Loading 𝑘𝐴 𝑚−1 400
Power factor − 1
Winding Factor − 0.9
Active Power Density 𝑘𝑔 𝑚−3 3000
Mean Stator Factor − 0.66
Cryostat Adding Weight Factor − 0.15
Rated Output Power 𝑘𝑊 1470
Rotational Speed 𝑟𝑝𝑚 11100
Length/diameter (L/D) − 1
4.3.1 The environmental screen
An environmental screen is required to contain the magnetic field within the
electrical machine space in order to screen the environment from stray
magnetic fields. Within a superconducting machine for a given field current the
magnetic flux density at the armature depends on the type of the environscreen
employed. The effect of this screen at the field winding is small since the
generator/motor is practically air cored, however at the stator which is closer
this effect is pronounced (Bumby, 1983). There are two main types of
environscreen normally being employed: the iron environmental screen and the
conducting screen.
Iron Environmental Screen
This type of screen is typically being chosen in applications such as power
station turbo-generators where the high power output per length and the low
98
screen losses make this type of screen an attractive option. The thickness of
this screen can be calculated by the following equation:
𝑡𝑐 =1
𝑝
�̂�
𝐵𝑚𝑎𝑥(
𝑟𝑠
𝑟𝑥)𝑝+1
2𝑟𝑥
1 + (𝑟𝑠
𝑟𝑥)2𝑝
𝑚 (4-15)
The thickness of the iron screen and consequently its weight depends on the
number of poles of the machine. The mass of this screen is being given by the
equation:
𝑊𝑠𝑐𝑟𝑒𝑒𝑛 = 𝛾𝜋 ∗ (𝑡𝑐2 + 2𝑡𝑐𝑟𝑥) 𝑘𝑔 𝑚−1 (4-16)
Equations (4-15) and (4-16) are being implemented in Simulink as an additional
subsystem to the previous machine models. The subsystem’s inputs and
outputs can be summarised in the following table:
Table 10 Inputs/Outputs of the Environmental Screen Subsystem
Inputs Units Outputs Units
Pair of poles (𝑝) - Screen Thickness (𝑡𝑐) 𝑚
Environscreen density (𝛾) 𝑘𝑔 𝑚−3 Environscreen Mass (𝑊𝑠𝑐𝑟𝑒𝑒𝑛) 𝑘𝑔𝑚−1
Mean stator radius (𝑟𝑠) 𝑚
Inner screen radius (𝑟𝑥) 𝑚
99
Figure 51 Environmental Screen Simulink Sub-model
Three different cases were investigated to demonstrate the extra weight that an
iron environmental screen will add to these machines: (a) a case where no
environmental screen is implied to the machines, (b) another case where an
iron screen is being used in a 4-pole machine and finally (c) a case where an
iron environscreen is added to an 8-pole superconducting machine. The last
two cases will show the dependence between the pole numbers and the mass
of the iron screen. A 4-pole and an 8-pole configuration were chosen as they
seem the most probable options for our application where frequencies between
400 and 800 Hz will be required (this is the case for Boeing’s 787 Dreamliner
aircraft).
100
Figure 52 Overall weight of a superconducting machine (a) without
environscreen, (b) with iron screen (4-poles machine) and (c) with iron screen (8-
poles machine).
It is clear that as the number of poles increases the weight of the iron
environmental screen drops significantly. However, even for an 8-pole machine
the added weight because of the iron screen is unacceptable becoming one of
the predominant machine parts on the weight estimation. Since it does not
seem likely that the superconducting machines will consist of high number of
poles, the additional weight of this type of screen clearly suggests that such a
screen will not be used at least in the machines which will be designed for
airborne applications. Particularly for high power machines (over 10MW) the
weight of the screen becomes prohibitive.
Conducting Environmental Screen
Another popular type of environmental screen is the conducting screen. Copper
and aluminium are the possible material choices for this type of screen. The low
density of aluminium (2700 𝑘𝑔 𝑚−3) compared to copper (8960 𝑘𝑔 𝑚−3) favours
0 20 40 60 80 100 120 140 160 180 2000
100
200
300
400
500
600
700
800
900
1000
Power Output (x100kW)
To
tal W
eig
ht (k
g)
no screen
iron screen (4 poles machine)
iron screen (8 poles machine)
101
aluminium screens in the machines designed for the airborne applications.
Furthermore, the weight of an aluminium screen is about the one tenth that of
iron screens (Bumby, 1983) and the following figure presents its effect on the
overall weight of the fully superconducting machines.
Figure 53 Overall weight of a superconducting machine (a) without
environscreen, (b) with aluminium screen (4-poles machine) and (c) with
aluminium screen (8-poles machine).
In this case the weight of the screen is not anymore the predominant weight
factor and particularly for an 8-pole machine it seems like an obvious choice.
However, this type of screen suffers from significant power losses which
degrade the overall efficiency of the machine. The eddy currents in this type of
screen affect the efficiency of the superconducting machines eliminating one of
the main advantages of these machines. In order to achieve a similar power
loss to the iron screen a substantially greater screen radius is required. Thus, a
compromise between the screen dimensions and its power losses is necessary.
0 20 40 60 80 100 120 140 160 180 2000
100
200
300
400
500
600
700
Power Output (x100 kW)
To
tal W
eig
ht (k
g)
no screen
aluminium screen (4-poles machine)
aluminium screen (8-poles machine)
102
Superconducting Environmental Screen
It is clear that, although for different reasons, both the iron screen and the
conductive screen could be a limiting factor for the overall attractiveness of
future fully superconducting machines. The weight of the former and the losses
added by the latter demonstrate the need for an alternative solution.
Unfortunately, there is no way of avoiding the environmental screen in this type
of machines. However, there has been a suggestion by some experts from
Rolls Royce that a superconducting environmental screen could be used
instead (patent pending). The thickness of this screen is expected to be
significantly low and so will be the added weight. The power density of the
superconducting machines will be affected but its effect was considered
negligible for the purposes of this study.
4.3.2 TRV Factor
In this study TRV factor is defined as the product of the magnetic and electric
loading of the electrical machines as shown in equation (4-17). Usually, the
designer aims to maximise the power output of a machine of given dimensions
by reaching the highest possible values of these two parameters. These
parameters present significantly higher operational limits in the SMs than in the
conventional machines. This is the main reason why these machines
demonstrate increased power density whilst the removal of Joule loss in the
excitation winding also increases their efficiency.
𝑇𝑅𝑉𝑓𝑎𝑐𝑡𝑜𝑟 = 𝐴 × 𝐵 𝐴 𝑚−1 𝑇 (4-17)
Generally, B is limited by saturation and iron losses whilst A is constrained by
the efficiency of stator cooling, by winding vibration, and by the space available
for the conductors (Miller and Hughes, 1977). Since there are not any data
available for fully SMs, their synchronous and transient reactances are
unknown. In order to be able to investigate their performance relatively reliably it
is reasonable to assume that the improvement rate of these two factors (i.e.
magnetic and electric loading) will follow similar trends. This is why these two
inputs were chosen to be studied together as a common “TRV factor” and its
103
effects on the total weight of the machines were investigated. The range of the
factor varied from 200 to 2200 𝐴 𝑚−1 𝑇 based on the expected maximum
values of these two parameters. There have been experimental work showing
electrical machines able to trap magnetic field in the area of 4T (Rada et al.,
2015) while this value could potentially reach even the maximum value of 10T.
Electric loading values around 700 kA/m have also been achieved (Tixador and
Daffix, 1997).
Figure 54 TRV Factor Vs. Total Weight of Fully Superconducting Machines
Quite similar trends can be seen on both versions with the first version (TRV
Original) showing more clearly the effects of this factor. It is obvious that after a
point the improvement of TRV factor does not reflect significant gains in the
total weight estimate, a result that we should take into consideration when
designing these machines. However, it is important to note that the rest of the
input parameters remained constant something that in reality will be extremely
difficult. Rotor and stator dimensions will have to change as well to maintain the
high values of magnetic flux. Furthermore, in the upper limits of this factor,
enormous electromagnetic forces should be expected and ways to control them
will be necessary. In our models a value of 760 was chosen based on the
PSAM results. This value seems to be the ideal based on this graph, a fact that
could be used as another validation for our results. In this graph, as well as for
0
20
40
60
80
100
120
140
160
180
0 1000 2000 3000
Tota
l We
igh
t (k
g)
TRV Factor (T kA/m)
Axial FieldConventional PMMachine
TRV Original
TRV optimistic
104
the rest of the sensitivity study, a comparison with the weight of a conventional
axial field PM machine (Jewell, 2009) of the same torque rating is also made.
4.3.3 Active Power Density
In this study active power density is defined as the average power density of the
active parts of the electric machines (i.e. stator and rotor). The effect of this
assumption on the total weight of the machines is very strong and an accurate
estimation of this value will be really crucial. Depending on the materials being
used as well as the machine architecture being chosen it is fair to assume a
range between 2000 and 8000 𝑘𝑔 𝑚−3. A value between 3000 and 4000
𝑘𝑔 𝑚−3seems more reasonable in the configuration under investigation in this
project. For the active parts of the machine, in the initial studies as described in
4.2.3 the stator is constructed mainly of epoxy concrete with a power density of
2400 𝑘𝑔 𝑚−3. The rotor on the other hand is constructed primarily of steel
(power density of 7859 𝑘𝑔 𝑚−3) with a small proportion of HTS material.
However, a hollow space of the same diameter as the shaft is being assumed
and this will significantly decrease the active power density, hence the
3000 𝑘𝑔 𝑚−3assumed value.
105
Figure 55 Active Power Density Vs. Total Weight of Fully Superconducting
Machines
The effect of the active power density on the total weight of the machines can
be seen in Figure 55. Although the trends of the two versions seem slightly
different, the truth is that in both cases the total weight was eventually
quadruplicate. However, even with the more pessimistic assumptions SMs are
still a lot lighter than the reference machine.
4.3.4 Cryostat Weight
The cryostat weight is another main assumption of the model. In order to
overcome this relatively unknown field, especially for the superconducting
machines, an expert’s opinion was asked. More specifically, Steven Harrison
(founder and former director of Scientific Magnetics-Oxfordshire) suggested that
a range between 10-50% of the active weight should be considered. However,
his view was that since we are interested in aerospace applications it is more
reasonable to look at the lower limits of this range. Thus, a sensitivity study for a
range between 8-30 % was carried out for both versions of the TRV models.
0
20
40
60
80
100
120
140
160
180
Tota
l we
igh
t (k
g)
Active Power Density (kg/m)
TRV Original
TRV Optimistic
Axial Field ConventionalPM Machine
106
Figure 56 Cryostat Weight Factor Vs. Total Weight of Fully Superconducting
Machines
The cryostat weight seems to have a stronger effect on the first version
however this is not entirely truth. The heavier the machines we investigate the
strongest the effect of the cryostat weight and that will be the case in both
versions.
4.3.5 Winding factor
In three-phase AC electrical machines the winding factor is responsible for the
decrease of the generated RMS Voltage. Most conventional machines have
winding factor values between 0.85 and 0.95 (Skaar, Krovel and Nilssen, 2006).
In our case a wider range between 0.80 and 0.97 was chosen and the derived
results assume that the rest of the parameters remain constant. In the next
graph as it was expected, it can be seen that the higher the winding factor value
the lighter our machines will be:
0
20
40
60
80
100
120
140
160
180
8 10 12 14 16 18 20 22 24 26 28 30
Tota
l We
igh
t (k
g)
Cryostat Weight Percentage (%)
TRV Original
TRV Optimistic
Axial Field ConventionalPM Machine
107
Figure 57 Winding Factor Vs. Total Weight of Fully Superconducting Machines
In PSAM Aerospace Assessment a value of 0.9 was assumed, however a
winding factor around 0.95 could be a reasonable assumption for these
machines. As it was expected an almost linear relationship was derived both for
the first and for the second version of the model. However, it is clear that the
winding factor does not play a predominant role in the overall weight of these
machines.
4.4 Key Model Limitations
This model is based on the TRV concept that is normally being used as a
preliminary sizing method for the design of conventional permanent magnet
machines. The main characteristic of this method is that it is relied on the
physical principles of the machines. Hence, some important limitations must be
borne in mind when using this mass estimation method:
Structural limitations were neglected: superconducting machines
demonstrate high values of electric and magnetic loading compared to
the conventional machines. This characteristic apart from bringing
significant benefits it also creates some structural considerations
because of the very high forces being anticipated. The expected flux
densities dictate the use of minimum iron within the structure of the SMs
0
20
40
60
80
100
120
140
160
180
0.8 0.82 0.84 0.86 0.88 0.9 0.92 0.94 0.96
Tota
l We
igh
t (k
g)
Winding Factor
TRV Original
TRV Optimistic
Axial Field ConventionalPM Machine
108
(Hughes and Miller, 1977) whilst for a machine of similar dimensions to a
conventional one, larger forces must be carried. This means that
especially for smaller machines the structural weight could potentially be
the dominant weight factor. Since this model accounts only for the
electromagnetic considerations, the calculated mass of small machines
appears low and must be treated only as an optimistic preliminary sizing
estimate.
Machine Losses not physically modelled: as an input in this model an
assumed efficiency is being used. This assumption was based on the
very optimistic NASA predictions for the N+3 timeframe, where efficiency
around 99.97% was predicted (Felder et al., 2011b). For the DEAP
project an assumed value of 99.9% (Wright et al., 2015) was used and
based on this value the thermal load of the machines was calculated.
However, electrical losses being produced because of electromagnetic
considerations were not calculated in the existing model. Operating
temperature and speed will most probably affect the overall efficiency of
these machines which are operating in cryogenic temperatures. These
inefficiencies are critical on deciding the economic feasibility of this
aircraft and hence higher fidelity models that take into account these
losses need to be developed in the near term future. It is also expected
that losses other than electrical and thermal can be kept at ambient
temperature, for example by designing the bearings outside the main
cryostat. The overall efficiency of the machines would suffer from some
bearing and windage loss too. This has not been considered in the TRV
model. Other design concerns, such as heat dissipation, may play a
more dominant role when these machines become small and hence it is
more likely that they will not scale linearly.
Overall weight estimation did not include all the parts of the
machine: in chapter 4.3.1 it has already been pointed out the lack of
environmental screen weight estimate for the overall weight calculation of
the SMs. This is expected to be a lightweight superconducting screen
where its weight will be negligible compared to the rest of the
109
components. Furthermore, in the existing model there is no weight
consideration for non-active machine parts such as the thrust bearings,
the vacuum chamber, the end flange and other structural support. The
latter might play an important role particularly in the smaller machines,
but in principle all these parts have a secondary effect on the overall
weight of electrical machines and at least for preliminary estimations can
be neglected.
4.5 Model Validation
There is no other method in the literature that estimates the weight of the fully
superconducting machines. Moreover, there are no data available for built fully
SMs which could have been used as a reference. The only possible comparison
could be made with the NASA TeDP concept machine weight estimations.
These are considered relatively optimistic and there is no background data
describing the methodology that was followed to derive these values. The
following table shows the weight estimates of the superconducting machines
derived from NASA N+3 predictions as well as by the two versions of this
research study. In the NASA study two 53khp, 6500 rpm superconducting
generators and fourteen 7.7khp, 4800 rpm superconducting electric motors
were used (Brown, 2011).
Table 11 Comparison between NASA and TRV model weight estimates
Machine Units NASA
(BSSCO)
NASA
(MgB2)
TRV
(Original)
TRV
(Optim.)
Generator (w/o cooler) 𝑘𝑔 684 949.3 1144.9 172.16
Motor (w/o cooler) 𝑘𝑔 196.4 225.4 252.5 40.1
The first thing to notice in the table above is the extremely low weight estimates
with the TRV optimistic version. This comparison definitely questions the
reliability of the optimistic version of the TRV method. Even looking on the
sensitivity study of these two methods, it was apparent that the optimistic
version in most occasions has given unrealistic results, giving more than an
110
order of magnitude lighter machines than the original version. On the other
hand, the original version seems somehow closer to the NASA predictions. In
general, all NASA predictions for this concept could be considered rather
optimistic and since the TRV (original) estimates are less than 20% heavier
than the NASA (MgB2) values it seems like this method could be considered as
a pessimistic prediction for the weight of the fully SMs. It is important to note
that NASA predictions vary with the superconducting material being used
(BSSCO or MgB2) and more specifically are a function of the superconductor
filament diameter. Clearly, a different approach has been followed in this study
and hence a reliable comparison cannot be made.
The TRV models of this research study were presented and used throughout
the DEAP project. Several experts from all the partners involved (i.e. AGI, RR,
CU, and Cambridge University) have validated the reliability of the TRV original
method, whilst the weight estimates were used in the overall DEAP system
weight estimation (Berg et al., 2015b).
111
5 Superconducting Electric Aircraft (SEA)
Up until this point the concepts of TeDP and SPN have been discussed. At this
stage the possibility of applying these approaches to simpler aircraft electric
systems will be investigated. The key example to this is the More Electric
Aircraft (MEA) and therefore a study has been carried out to examine the
possibility of using SPNs in such a system. The Boeing “Dreamliner” 787 and
the Airbus 380 aircraft have been the first MEA in use. Both airplanes prove to
be remarkably successful but there are reasons to believe that the industry is
currently facing some important obstacles on scaling the aforementioned
reference aircraft. Superconductivity could potentially solve most of these
scaling issues, particularly if it will be combined with the TeDP concept where a
superconducting network will also be present.
In this chapter a brief description of the MEA concept will follow, the electrical
power network of the 787 aircraft will be described and a comparison between
the use of a conventional and a superconducting network for the electrical
system of such an aircraft will be made. The study will then be extended to
different sized aircraft examples so that a wide range of electrical load demand
could be investigated.
5.1 More Electric Aircraft (MEA) Concept
5.1.1 MEA Concept Description
The aviation industry was always driven by the demand to optimise aircraft
performance, whilst reducing the operational and maintenance costs and
increasing the reliability of the whole aircraft. In the last few years an extra
objective to provide some more environmental friendly solutions has pushed
toward a more electric approach in the design of current and future airplanes. A
MEA it is typically characterised by the extended use of electrical power in the
Secondary Power System (SPS).
These systems form the non-propulsive parts of the aircraft and in conventional
configurations are driven electrically, mechanically, hydraulically or via
112
pneumatic/bleed air power (Laskaridis and Pilidis, 2004). Figure 58 (Jones,
2002) demonstrates the SPS of a traditionally powered aircraft. In such an
aircraft, pneumatic power is obtained from the main engines’ High Power
Compressor (HPC) to power the Environmental Control System (ECS) and to
provide hot air for the Wing Anti-Icing (WAI) System. Mechanical power is
transmitted via gearboxes from the engines to central and local hydraulic
pumps, to the main electric generator as well as to other mechanically driven
subsystems. On the other hand, the actuation systems for primarily and
secondary flight controls mainly use hydraulic power. The same goes for the
landing gear and other ancillary systems. Finally, electric power derived from
the main generator powers the avionics, the cabin power demands (lights,
galley, in-flight entertainment etc.) and the aircraft lighting (Rosero et al., 2007).
This combination of secondary power types has always being debated because
of the additional complication and the resulted reduced efficiency of the overall
system efficiency (Abdelhafez and Forsyth, 2009).
Figure 58 Conventional secondary power systems (Jones, 2002)
In the MEA concept electric power becomes the main way of distributing power
to the majority of SPS. The expanded electric network now also includes the
cabin pressurisation system, the ECS, the WAI, flight control actuation, landing
gear, doors, fuel pumps and engine’s ancillaries.
The motivation behind the more electric approach is the reduction of the
operating costs, the decrease in fuel burn and last but not least the limitation of
the environmental impact. In a MEA the hydraulic system is removed leading to
113
a reduced system weight and the simplification of the maintenance procedure.
The reliability, vulnerability and redundancy of the aircraft are also improved
without the presence of a complex hydraulic subsystem. Moreover, the
elimination of the pneumatic power significantly improves the efficiency of the
“bleedless” main engines. Much better cabin environment for both the
passengers and the crew is possible in such an arrangement whilst the aircraft
fuel burn is also reduced. Many heavy engine components such as bleed
ducting, pre-coolers and ECS, which used to cool and pressurise engine offtake
air, will no longer be needed (Provost, 2002). It is clear that a More Electric
Engine (MEE) seems ideal in a configuration such as the MEA and many
studies across Europe and worldwide have been focused on similar engines
(Hirst et al., 2011). Figure 59 shows a comparison between a conventional and
a MEA aircraft system.
Figure 59 Comparison between conventional and MEA systems (Provost, 2002)
As indicated earlier three different individually optimised subsystems will be
replaced by a common electrical system that will control the majority of SPS
functions in a MEA. The main challenge is the optimisation of this electrical
system, where a trade-off between AC and DC systems as well as voltage and
current levels of the whole power network is necessary. In this sensitivity study
the proposed superconducting concept that will later be described (5.2.2) could
solve most of the issues presented in a MEA.
114
5.1.2 Airbus 380
The first aircraft with some more electric characteristics that entered the
commercial service in 2007 was Airbus 380 (Figure 60). The main electric
novelty in this aircraft can be found in the flight control architecture. Traditionally
three hydraulic systems produce the required power for the flight controls.
However, in the case of A380 the additional weight and complexity of the
required hydraulic system due to the larger dimensions of the plane made this
option particularly unattractive. Instead, hydraulic combined with electric flight
control architecture was preferred. Hydraulic power is still the main power
source for the flight controls, but many electrically powered actuators were used
in order to save weight and reduce the complexity of the system.
Figure 60 Airbus 380 aircraft (image courtesy of Airbus)
Overall benefits such as improved maintainability and reliability as well as
reduced weight and cost were considered significant innovations for the A380
aircraft. Safety margin was also increased due to the use of different power
sources (Adams, 2001). .
A380 power system distribution can be seen in Figure 61. It consists of a
primary 400VAC power bus (doubling the one of previous systems) with a
variable frequency between 360-800 Hz. The Variable Frequency (VF) power
generation enabled the reliable production of additional power with extra weight
and maintenance costs benefits compared to previous systems (Adams, 2001).
115
Figure 61 A380 Power distribution system (Abdel-Fadil, Eid and Abdel-Salam, 2013)
A similar power distribution system is also used to the second and most
representative example of a MEA currently in service. This would be the Boeing
787 “Dreamliner” aircraft which will also be used as a reference for the
upcoming proposed architectures.
5.1.3 Boeing “Dreamliner” 787
One of the most successful and popular aircraft nowadays, with more than a
thousand orders already, is the Boeing Dreamliner 787 (Figure 62). This is the
closest example to the MEA concept. The main difference between 787 and
other more conventional models is its emphasis on electric systems, which are
aiming to replace most of the existing pneumatic subsystems in the traditional
architectures.
116
Figure 62 Boeing Dreamliner 787 Aircraft (image courtesy of Boeing)
The Boeing 787 is without any doubt on the cutting edge of airliner innovation,
being considered by many experts as an “ahead of its time” aircraft (Thiétart,
2013). It was specifically designed to be 20% more fuel efficient than the 767
model, with its electrical system being the major component that made this
aircraft so innovative. On the 787 the only remaining bleed system is the anti-
icing system for the engine inlets. The whole system architecture is completely
changed with systems such as the pneumatic engine, the APU start motors
and load compressors, pre-coolers, various ducts and air control systems being
just a few of the eliminated components of this novel aircraft (Hale, 2008). This
transition from bleed-air to electrical power significantly reduces the complexity
of the mechanical system in 787. The mechanical complexity of braking is also
reduced by the use of electrical, instead of hydraulic, actuators. Leak and
overheat detection systems for hydraulic fluid leaking are no longer needed,
whilst failures of electric brake actuators could easily been handled without
severe performance penalties. Overall, mechanical complexity has been
reduced up to 50% compared to a 767. This development has reduced
accordingly the maintenance costs, while the system’s reliability is increased
with improved health monitoring and fault tolerance (Hale, 2008).
Figure 63 (Sinnet, 2008) presents a general overview of the electrical system in
a 787 aircraft compared to a similar size conventional aircraft architecture. The
Dreamliner’s architecture consists of six generators (2x 250kVA per engine and
117
2x225kVA for the APU) operating at 235 VAC. These generators are directly
linked to the engines’ gearboxes operating at a variable frequency of 360 to 800
Hz depending on the speed of the engine. The electrical system features one
forward and one aft electrical/electronics (E/E) bay, as well as a number of
remote power distribution units (RPDU) for supporting airplane electrical
equipment. The system saves weight by reducing the size of power feeders.
The system also features two forward 115 VAC external power receptacles to
service the airplane on the ground without the APU and two aft 115 VAC
external power receptacles for maintenance activities that require running the
large-rated adjustable speed motors. All the aforementioned subsystems can be
seen in Figure 63.
Figure 63 787’s electrical system compared to traditional aircraft architecture
A more detailed description of the electric power network of this aircraft will
follow in the next section since this system will be used as a reference for the
proposed design architecture.
5.1.4 Going Beyond 787: Challenges and design options
The “Dreamliner” aircraft have showed some significant benefits derived from
the MEA concept making it the most popular model nowadays. Ideally, this
118
approach would be extended to aircraft of different sizes whilst the extent of
electrification will continue to grow. However, several technical reasons make
the scaling of MEA approach problematic.
Electrical power systems present a number of disadvantages when applied to
aircraft applications mainly due to the adding weight. Electrical machines and
power electronics could be rather bulky especially for high power applications.
In larger aircraft than the 787 the cables’ size becomes another obstacle. In
order to deal with the increased electric loads, there are two design paths:
either to increase the current levels of the distribution network or to increase the
system voltage levels. The former increases the weight and volume of the
power cables whilst the latter suffers from the corona discharge effect. Another
issue of high power networks is the efficiency of the whole network. In 787 there
is an electric power load around 1MW. This number could be increased with the
further electrification of future aircraft and/or with the use of MEA concept to
longer range aircraft. The efficiency of a typical electrical system could be in the
range of 97-98%. While this efficiency might seem high, it could give several
hundreds of Watts heat losses which will create thermal management
considerations. Finally the potentially higher currents of the system also
increase the fault currents of the network. Further heavy protection and
switching equipment will be necessary to deal with these fault arcs. The
motivation behind this chapter is to address all these issues for next generation
aircraft. The author believes that superconductivity could be the main enabler
and problem solver for the majority of MEA challenges.
As described earlier (Chapter 4) fully superconducting machines will be
significantly lighter and will occupy less space than the conventional equivalent
machines. At the same time, superconducting cables’ current capability is
expected to eliminate the high currents constraint in a conventional power
network. The derived fault currents will now be controlled by the SFCLs
(Chapter 3), while no mechanical switching will be used and superconducting
equipment will also be used both for switching and protection devices.
Furthermore, the efficiency of the superconducting network is going to offer
119
further benefits. Both the machines and the cables are expected to be more
than 99.9% efficient solving the thermal management problem. Regarding the
power electronics, early studies have shown that operating them in cryogenic
temperatures improve both the efficiency and the power density of these
devices (Leong, 2011). A more detailed description of the performance of
“cooled power electronics” will follow. It is clear that the use of a
superconducting network will complicate the electrical system adding another
important secondary system (i.e. cryogenic cooling system), but the author
believes that the added complexity will be compensated by the numerous
advantages that such a configuration could offer.
In the next subchapters, by using as a reference the 787 electrical system a
comparison between the use of a superconducting and a conventional electrical
network will be made. After that, different aircraft sizes and electric load
requirements will be explored aiming on identifying the areas and limits where a
superconducting solution could be proved beneficial in terms of weight.
5.2 Superconducting Electric Aircraft Approach
This chapter will start with a detailed description of the loads and components of
the electrical power network of the 787 aircraft. Using as a baseline the exact
same power network, a superconducting version of it will show any derived
benefits and/or constraints. The conclusions of this comparison will be used for
a broader sensitivity study where different electric power levels will be explored.
This study will mainly focus on the weight of the secondary power network in
both the conventional and the superconducting case. Four different cases will
be explored based on the assumptions being made. There will be two cases for
conventional networks (current and future values) and two for the
superconducting versions (NASA and DEAP assumptions).
5.2.1 787 Electrical System Overview
A brief description of the secondary power system of the 787 model has already
been presented. In this subchapter a more detailed representation of the
electrical power network, which will include the secondary loads, the various
120
components (such as the electrical machines and the power electronics) and a
general overview of the whole architecture will be described. The efficiency and
weight of the main components will be given, whilst these values will also be
used as a reference to the next subchapter where a superconducting version
will be analysed.
Electric power demand in an aircraft changes depending on the flight phase.
Typically, the Top of Climb (TOC) phase is the most demanding for the
electrical power network but practically there are no significant differences in the
total power demand during the whole flight mission (only the individual
secondary loads change). More specifically, ice protection and hydraulics might
require more power at lower altitudes, while ECS and cabin pressurization could
be the dominant loads during cruise (Whyatt and Chick, 2012). However, the
demanded electrical power drawn from the engine generators remains relatively
constant throughout most of the flight mission. Figure 64 demonstrates the
electrical power demand from the 787 main generators during major flight
phases.
Figure 64 Total Electrical Power Demand during several flight stages of the 787
aircraft (Whyatt and Chick, 2012)
As it can be seen, the total electric load remains relatively constant at a value
slightly lower than 1MW. For simplicity reasons, it has been assumed that a
121
constant power demand is kept during the whole flight mission and for the
following analysis cruise is chosen as the design point.
Another important parameter of the upcoming analysis is the distribution of the
several secondary loads. In 787’s power network there are four different
distribution buses: a main 235 VAC distribution line, a +/-270 VDC line which
includes important loads such as the ECS, an 115 VAC 400 Hz bus for loads
such as the ICS and finally a +/- 28 VDC transmission line for a smaller portion
of secondary loads. It is clear that such a power network requires a significant
number of power electronics. It is also a fact that the efficiencies of the electrical
components in the 787 model are significantly improved compared to previous
non-MEA aircraft (Whyatt and Chick, 2012). Advances in electronics and the
use of Variable Frequency Starter Generators (VFSGs) are the main reason for
the increased efficiency and power density of the electric components. The
most important distributed loads and the efficiencies of the several components
of the 787 case are summarised in the following diagram.
Figure 65 Electric loads and efficiencies diagram of the 787 electrical power
network
122
The Boeing 787 aircraft is typically powered by the RR Trent 1000 engines. The
Trent 1000 engine is a three shaft high bypass ratio turbofan that was
specifically designed to power the 787 aircraft. It is a bleedless engine designed
to fit the MEA concept requirements enabling increased levels of electric power
to be transferred through the Intermediate Power (IP) spool. The engine has
been up to 12% more fuel efficient than the previous model of Trent family (i.e.
Trent 800) whilst there is 40% less emissions than the current legislation
requirements (Ojha and Raghava, 2014). In this study a fuel efficiency of 59%
is used (Whyatt and Chick, 2012).
Figure 66 Trent 1000 three shaft configuration (Ojha and Raghava, 2014)
The electrical system generates the required power by extracting mechanical
power from the engine accessory gearbox. In the 787 case two generator pads
(2x250kVA) are provided-the term pad describes the part where a mechanical
device mounts on the gearbox. Due to the higher frequency required from the
VFSGs compared to the constant frequency 400 Hz AC power of the
conventional integrated driver generators, the pad speed of the 787 accessory
gearbox is higher (Moir and Seabridge, 2013). A typical efficiency of 97% was
chosen in the sensitivity study of this chapter.
As it has already been mentioned, 787 incorporates four 250 kVA VFSGs
(Figure 67) connected directly to the engine gearbox. The generated frequency
of these machines depends each moment on the speed of the engine. On the
123
ground these machines-powered by the APU- are used to start the engine,
whilst when the engines are running the VFSGs are the primary source of
electric power (Boeing, 2013). This type of power generation is considered as
simpler and more efficient since it does not involve the complex CVG (Constant
Velocity Gearbox) or IDG (Integrated Drive Generator) subsystems (Moir and
Seabridge, 2013). Hence, the reliability increases and the maintenance costs
fall accordingly. The main benefit delivered by the implementation of this
configuration is the elimination of the bleed system which was typically used to
feed the ECS system. Heavy bleed air components are no longer present
enabling significant weight savings, while the elimination of the energy losses of
the bleed air system enhances the efficiency of the overall electrical power
network. An overall efficiency of 92% was assumed for the purposes of this
project.
Figure 67 Variable Frequency Starter Generator (VFSG) used in 787 (Clark, 2012)
One of the most important components of the network under investigation is the
power electronics being used. Compared to Boeing 777 aircraft, the 787
involves the use of state-of-the-art highly efficient power converters
manufactured by Thales Group (THALES, 2015a). More specifically, an Auto
Transformer Rectifier Unit (ATRU) converts +/- 230 VAC to 270 VDC in an
efficiency that reaches values over 97% (THALES, 2015b), while another
transformer achieves 98% efficiency in converting power from +/- 230 VAC to
+/- 115 VAC (THALES, 2015c). On the other hand, a less efficient Transformer
Rectifier Unit (TRU) is being used to convert +/- 230 VAC to 28 VDC where
124
efficiency up to 85% can be reached. These high power conversion efficiencies
had an important positive effect to the attractiveness of the MEA concept.
The main electrical loads of 787 are presented in Figure 65. Almost half of the
required electrical power during cruise comes from the +/- 270 VDC distribution
bus with the ECS load to be the most demanding load of the aircraft (~320 kW).
Another relatively demanding load is the wing anti-icing system which in 787
requires in the order of 100kW of electrical power (Moir and Seabridge, 2013).
Moreover, the electric motor pumps which replaced the traditional hydraulic
engine driven pumps require around 400kW in total. The electrically powered air
conditioning packs are located in the central sector of the aircraft, whilst the
engine starter motors and the electric motor pumps are mounted in the left aft
distribution panel of the plane. A more detailed representation of the 787
topology could be seen in Figure 68.
Figure 68 Electrical Power Distribution System in 787 (Moir and Seabridge, 2013)
5.2.2 Superconducting Version of 787 Electrical Power Network
In this subchapter a superconducting modification of the 787 electrical system
will be investigated. This version will include fully superconducting electrical
machines, cryo-cooled power electronics and a superconducting distribution
system. The required weight and power for the cooling system necessary for
this version will also be calculated using cryo-cooler models being developed
during the DEAP project and were validated by several experts during this
125
project (Berg et al., 2015b). A brief description of the chosen cryo-cooler model
and the main assumptions being made will be described in Appendix A.2. Two
different superconducting versions will be explored: one based on the
component efficiency and weight assumptions made during the DEAP program
and another one based on the assumptions that NASA has used in their
sensitivity studies for the TeDP concept.
Electrical machines
The ‘Dreamliner’ aircraft, as it has already been described, incorporates four
main VFSGs and two back-up APU generators. The benefits of the variable
frequency system have been pointed out in the previous subchapter. Even with
the improvements in efficiency that such a system introduced, fully
superconducting machines will still be significantly more efficient.
With the current technology standards fully superconducting machines seem
possible only with the use of MgB2 material for the stator of these machines.
This leads to an operational temperature of 20K for the electrical machines.
However, their efficiency even for the pessimistic DEAP case will reach a value
of 99.9% which is more than 7% improvement than the currently used VFSGs.
Their weight can be estimated using the models presented in Chapter 4.2. Note
that the nominal power of these machines needs to be increased to
compensate for the additional power requirements of the cooling system. Even
with such an addition however both the weight and the efficiency of these
machines will be enormously better.
NASA has claimed the feasibility of constructing fully superconducting machines
which will be 99.97% efficient. This value has been used for the second case of
this study. NASA has also investigated the possibility of using BSCCO as the
main stator material for their machines. This was based on the fact that future
improvements in HTS materials such as BSCCO and YBCO could allow the
production of these materials in round wire form rather than in the tape form that
are currently being produced. This will decrease their AC losses and will make
them clear favourite candidates for the future fully superconducting machines.
126
For these initial stages of the study only the engine generators were assumed
to be superconducting. However, a configuration where also the APU
generators are fully superconducting should be considered.
Superconducting distributed transmission lines
The distribution system of the 787 incorporates four different distribution buses
each dedicated for a number of secondary loads. The benefits of using
superconducting cables are numerous and have been pointed out in several
studies (Jin, 2007). In this system, the improved efficiency and most importantly
the significantly smaller conductor size is what makes their case more
attractive. Scaling up the 787’s electrical power network will have a direct effect
on the overall electrical power load. Having cables capable of carrying high
currents will both result to high losses and heavy transmission lines. Increasing
the voltage level of the whole network will most probably create corona effect
issues as it was pointed out earlier in this thesis. Superconducting cables can
solve all these issues. Especially for the DC buses (i.e. +/-270 VDC and +/-28
VDC) superconductors show no ohmic resistance reducing the distribution loss
and hence the thermal load of the network.
NASA has assumed typical losses on the order of 5 W/m of cable length (Xi et
al., 2006). This assumption was also used during the DEAP program as a
conservative prediction. Note that for all the superconducting cases MgB2 was
chosen as the primary material being used for the transmission lines (T=20K),
although there have already been applications where BSCCO transmission
lines were operating efficiently (Maguire et al., 2007). However, the 20K was
chosen as the most conservative case and as a way to keep the same
operational temperature for the whole system. The latter will simplify the cooling
system architecture and design/modelling process. For the weight predictions of
this study assumptions used during DEAP project were also the baseline for this
investigation. Power losses of 5 W/m and 8 W/m were assumed for DEAP and
NASA cases respectively. Note that during the DEAP project even more
optimistic predictions were claimed (Wright et al., 2015) but for the purposes of
this study a more moderate approach was chosen.
127
Figure 69 Boeing 787 and Airbus A350 size (www.AviationExplorer.com, 2015)
Figure 69 demonstrates the dimensions of Boeing 787 compared to the newest
Airbus model of similar range A350. In order to calculate the expected weight
and losses of the distribution system, an approximation of 75m of cable for the
main distribution lines was chosen based on the aircraft’s length and the
electrical power network architecture (Figure 68).
Cryo-cooled Power Electronics
It is generally noted that certain semiconductor materials become increasingly
efficient at cryogenic temperatures. It is true however that there have not been
many published studies on the topic of using cryo-cooled power electronics in a
superconducting system.
A DARPA (Defence Advanced Research Projects Agency) project focusing on a
system-level theoretical study to investigate effects of using cryogenics in the
power conversion components of a superconducting system was carried out
between 2002 and 2005 (Hennessy, 2009). Some of the study’s findings were
the significant benefits with respect to energy dissipation, access to higher
operating frequencies and improved reliability that many silicon-based power
electronics at cryogenic temperatures presented.
128
Generally, exact weights for the converter, as well as for the rest of the
components in such a configuration, are difficult to be estimated. Early in the
design phase is really hard to find all the necessary information to reliably
model the size of the various components. The simplest way to size
components is by using their anticipated power to weight ratios. This is the
method that was used for the weight estimation of cryo-cooled power
converters. This seems appropriate for this research study since it is a common
tactic especially for technologies of low TRL level, where conceptual designs
are investigated.
To determine the proper power to weight ratios for the power converters, firstly
the state of the art in high power low weight power electronics was examined. In
the conventional case the THALES product specifications were used since they
are the converters currently being used in the 787 aircraft. A power to weight
ratio up to 5 kW/kg could be observed (THALES, 2015d), whilst efficiencies up
to 98 % have also been achieved (Furmanczyk, 2009). Higher power density
values have been achieved in non-aerospace applications. For example,
manufacturers of inverters for electric cars have claimed current power to
weight ratios up to 10 kW/kg with expectations of reaching values around 15
kW/kg in the near term future (Rogers, 2012). It is reasonable to expect similar
values for the aerospace applications and this can be further enhanced with the
additional benefits of a cryo-cooled case. Therefore, a conservative assumption
of up to 20 kW/kg was used for the DEAP case. On the other hand, NASA has
been relatively optimistic in terms of power density of cryo-cooled power
electronics. A mass-specific power of 20 hp/lb (i.e. 32.8 kW/kg) and a 99.8%
efficiency without cooling has been chosen as a target for the 2035 timeframe
based on an unpublished report by MTECH Laboratories (Brown, 2011).
It is not clear yet which is the optimum cryogenic operational temperature for
the various power electronic devices. The findings of a PhD thesis (Leong,
2011) focusing on the use of power devices below 100K to minimise the power
losses can be summarised in the figure below. It can be seen that the optimum
range of operation varies for the different semiconductor materials. It is also
129
clear that lower temperatures do not necessarily have a positive effect on the
on-state behaviour of these materials with the example of Si n-channel and Si
p-channel MOSFETs performing better in temperatures between 50 and 100K.
However, for the purposes of this study an operational temperature of 20K was
used for the pessimistic case, so that the worst case scenario in terms of
cooling power demand to be explored. NASA has used as an operational
temperature both the MgB2 case (i.e.20K) but also a more optimistic
temperature of 111K without differing the power to weight ratio of these devices.
Figure 70 Tested behaviour of power electronic devices at cryogenic
temperatures (Leong, 2011)
5.3 MEA and SEA Weight and Efficiency comparison studies
(based on the Boeing 787 aircraft)
In this chapter a comparison between the weight and efficiency of the main
components of the secondary power network of the 787 aircraft and a
superconducting modification of it will be demonstrated. The components that
were included in this study are the electrical machines, the power electronics,
the main transmission lines and the required cooling system of these
configurations.
130
5.3.1 Basic Assumptions
Four different cases were investigated based on the optimism of the
assumptions being used. As a first case the reference aircraft (i.e. Boeing 787)
was used with the minimum assumptions possible since a lot of information for
the actual products was found in the literature. Moderate future predictions for
the power density and efficiency of the conventional 787 secondary power
system’s components were the baseline for the second case under
investigation. The other two cases are related to the superconducting proposed
architecture. DEAP project conservative estimates were used for the first
superconducting case, whilst NASA’s optimistic predictions for the 2035
timeframe were used as the second superconducting case.
It is important to summarise the assumptions being made for each case
separately concerning the power density, efficiency and operating temperature
of the various components. The latter two are the decisive parameters for the
weight estimation of the cooling system in the superconducting cases. Following
the structure of the previous subchapter first of all each main component will be
explored and compared separately and then a combined comparison study will
conclude the first stage of this study.
VFSGs
In this study only the weight and efficiency of the VFSGs was investigated
without taking into account the APU electrical machines. For the conventional
case the total weight of the four VFSGs was found to be around 363kg with
92% efficiency. For the conventional future 787 type of aircraft high speed
electrical machines with power density of 10 kW/kg and an efficiency around
98% was assumed based on a study that Airbus and AGI had carried out in
terms of radical aircraft concepts for a technology level beyond 2030 (Barraud
et al., 2015). In the superconducting cases a different rating of these machines
was necessary to counter for the extra power needed to drive the cryo-coolers
of these architectures. A rating of 350 kW was chosen securing 400 kW power
available to drive the cooling system. The cooling power demand was estimated
to be around 220 kW for the DEAP case and around 70 kW for the NASA
131
estimates, but a conservative approach was chosen to secure the reliability of
the results. This cooling power demand was also calculated using the double
stage reverse Brayton Cryo-coolers’ models developed for this study (A.2). The
weight of the superconducting electrical machines was calculated using the
models presented in 4.2.4 for the DEAP superconducting case, whilst for the
NASA case the torque density of the MgB2 superconducting motor of the N3-X
BWB aircraft (Brown, 2011) was used. Table 12 summarises all these
estimations:
Table 12 VFSGs key variables values for each case
Variable Units 787
Current
787
Future
Superconducting
Case (DEAP)
Superconducting
Case (NASA)
Rating 𝑘𝑊 250 250 350 350
Unit Weight 𝑘𝑔 90.75 25 19.78 9.16
Total Weight 𝑘𝑔 363 100 79.12 36.64
Efficiency % 92 98 99.9 99.97
Operational
Temperature
K Ambient Ambient 20 20
Power Electronics
The power electronics weight estimation is a complicated procedure. All the
main assumptions for the power converters of the system were mentioned in
the subchapter 5.2.2. However, it should be noted that these assumptions
should stand only for the ATRUs (+/- 230 VAC to 270 VDC) and the
transformers (+/- 230 VAC to +/- 115 VAC) of these electrical power networks.
In the conventional case, the TRU (+/- 230 VAC to 28 VDC) has a significantly
lower efficiency than the other two converters (i.e. 85% instead of 98%) and a
power density of only 0.65 𝑘𝑊/𝑘𝑔 instead of 5 𝑘𝑊/𝑘𝑔 . Hence, an efficiency of
90% was assumed for the 787 future and the DEAP superconducting cases,
132
whilst efficiency around 95% was used in the NASA case. Concerning the
power to weight ratios of the TRUs in each case, values three times lower than
the maximum expected power densities were used as an approximation.
Table 13 Power electronics key variables values for each case
Variable Units 787
Current
787
Future
Superconducting
Case (DEAP)
Superconducting
Case (NASA)
Power to
Weight Ratio
𝑘𝑊
/𝑘𝑔
Up to
5
Up to
15
Up to
20
Up to
33
Total Weight 𝑘𝑔 243 48.79 36.7 21.84
Efficiency % Up to
98
Up to
99
Up to
99
Up to
99.8
Operational
Temperature
K Ambient Ambient 20 20
Cables
The weight of the main cable span of the 787 aircraft was not available in the
literature. Instead the capabilities of the conventional cabling and the resulted
mass were based on existing manufacturer’s data; this data meets current
copper cable sizing practices. The maximum current density of a copper or
aluminium wire is limited to 4 𝐴/𝑚𝑚2 (Xi et al., 2006). Based on that value the
required cross sectional area of the 787’s main power cables was estimated to
be 1086 𝑚𝑚2. Data for copper cables of such a wide cross sectional area were
not available. Instead, it was possible to develop a relationship that links the
cross sectional area of the copper cable with its weight per meter.
𝐶𝑜𝑝𝑝𝑒𝑟 𝐶𝑎𝑏𝑙𝑒 𝑊𝑒𝑖𝑔ℎ𝑡 = 0.0129 ∗ 𝐶𝑟𝑜𝑠𝑠 𝑠𝑒𝑐𝑡𝑖𝑜𝑛𝑎𝑙 𝑎𝑟𝑒𝑎0.9593 𝑘𝑔/𝑚 (5-1)
This relationship (5-1) corresponds to PVC insulated stranded copper cables
and is based on data available in (Keison, 2014) and are presented in Figure
133
71. PVC insulated cables have been approved for aircraft use by the civil
aviation authority ((CAA), 2002).
Figure 71 Weight per meter of conventional copper cable with PVC insulation
In the future technology estimations a 10% improvement was assumed. This
conservative assumption was based on the fact that only improvements in
insulation could be achieved in this type of cable and typically the conductor
weight is what sizes the cables.
For the superconducting cases a 5 kg/m weight per unit length was assumed in
the DEAP case (Wright et al., 2015) , whilst NASA based on a study by Xi (Xi et
al., 2006) in their initial estimations assumed a value of 9.2 kg/m for the weight
of the main transmission lines. However, a weight approximation value of 500
A/kg/m has been used in a later NASA study (Armstrong et al., 2012) and that is
the value also being used in this study. It should be noted that in their initial
assumptions NASA was using as a reference a High Voltage superconducting
cable that requires a substantial level of insulation and this is not the case with
the power system under investigation. The heat losses of a superconducting
cable were assumed to be 5 W/m for both superconducting cases (Brown,
2011). In the conventional cases these losses were assumed negligible
y = 0.0129x0.9593 R² = 0.9965
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
0 50 100 150 200 250 300 350 400 450
Wei
ght
per
met
er (
kg/m
)
Cross sectional area (mm^2)
Copper Cable Weight
134
compared to the losses of the rest of the components and hence no assumption
was made.
Table 14 Main Cable line key variables values for each case
Variable Units 787
Current
787
Future
Superconducting
Case (DEAP)
Superconducting
Case (NASA)
Weight per
unit length
kg/m 10.55 9.5 5 8
Total Weight 𝑘𝑔 791.25 712.5 375 600
Losses W/m - - 5 5
Total Length m 75 75 75 75
Operational
Temperature
K Ambient Ambient 20 20
Cooling system
The cooling system has been characterised by many as the main drawback of
having a SPN in an aircraft. Complexity, reliability, and mainly extra weight are
some of the attributes that such a subsystem will add to the network. In the
conventional system there is also a cooling mechanism typically consisted of a
fully integrated package of pump, motor, controller, filter, and reservoir. It is not
an easy task to predict the weight of this system and since no data was
available in the literature for the 787 aircraft, an approximation was made; this
would be that the conventional cooling system weighs 30% of the overall weight
of the cooled components (Malkin and Pagonis, 2015). The same goes for both
conventional cases with the 787’s future case being more beneficial due to the
expected reduced weight of the components.
In a SPN, different options for the cooling system have already been described
(Chapter 2.3). In this study the option of cryo-coolers has been selected based
on the fact that extensive cryo-cooler studies have been carried out during the
135
DEAP program. The models that have been developed to estimate the weight
and the cooling power demand of the cooling system are described in detail in
Appendix A.2. For both superconducting cases Reverse Brayton cryo-coolers
were used (Palmer and Shehab, 2015). The same type of cryo-coolers was also
assumed during the NASA studies (Felder et al., 2011a) . In this study double
stage coolers were chosen for the configurations under investigation.
Compressor polytropic efficiencies of 90% and turbine polytropic efficiencies of
92% were considered as realistic assumptions for these models.
Superconducting motors were also used instead of conventional machines,
whilst a 5% drop in every heat exchanger of the system was also assumed.
These were the main assumptions of the models, but a full list of the
parameters being used can be found in Appendix A.2.
5.3.2 Results and Comments
Based on the aforementioned assumptions a weight comparison between the
different cases was carried out. These cases were considered as the most
representative to determine the feasibility and attractiveness of the SEA
concept. Both the DEAP program and NASA could be considered as reliable
references. The former investigated in depth the aspects of having a SPN in a
HEDP aircraft, whilst NASA is consistently working on the next generation
aircraft where superconductivity and all its derivatives are holding a significant
share on their research agenda. In regards to the conventional cases, it seemed
reasonable to use the current MEA aircraft in service (i.e. Boeing 787) as the
first reference since most of the system’s weight could be accurately found.
However, in order to make a more representative comparison it was necessary
to estimate the weight units by expressing the density (in terms of power,
torque, current or energy) of each component of the power network in the
timeframe that this future aircraft could move into production. Hence, the
second conventional case predicts the weight of the components for the 2035
timeframe. Figure 72 summarises the results of the first stage of this study:
136
Figure 72 Weight Comparison between different 787 MEA and SEA
configurations
As it was expected the superconducting electrical machines were found to
weigh almost five times less-with the DEAP estimates-than the current VFSGs
in use, whilst NASA estimates give more than an order of magnitude lighter
machines. The 2035 estimates for conventional electrical machines however
compensate somehow that gap by predicting rather competitive machines in
terms of weight.
A similar trend can be noticed in the case of power electronics. The expected
reduction in the size of the passive parts of the cryo-cooled power converters
will give significant weight benefits compare to the state of the art products.
More specifically, the power density of these components is expected to be
three (DEAP prediction) to ten (NASA estimate) times better in the
superconducting cases. However, once again, anticipated improvements in the
current power electronics’ technology could produce competitive products for
the 2035 timeline.
In regards to the cables’ weight in these configurations there are a lot of
remarks that need to be made before commenting on the results. First of all, the
presence of a SPN would most probably change the voltage and current levels
of the whole system. This will have a direct effect on the weight of all the
0
200
400
600
800
1000
1200
1400
1600
1800
VFSGs PE Cables CoolingSystem
Total
Wei
ght
(kg)
Conventional case
Future Conventional
Superconducting case (DEAP)
Superconducting case (NASA)
137
components, while in the case of cables this effect will probably be more
enhanced. In this study, for the conventional examples the main transmission
lines’ weight was calculated by using current aerospace manufacturer
datasheets and assuming a 10% improvement for the future estimates. This
improvement rate although seem moderate is only related to the insulation
technology development and not the conductor’s weight which is not expected
to alter. Contrary to the rest of their estimates NASA has been relatively
conservative on the predicted current density of superconducting cable. This is
partly due to the different nature of application their reference cables are built
for. The high voltage transmission line that has been used as a reference (Xi et
al., 2006) will require significant amount of insulation with capability of
withstanding voltages of well over 100kV, whilst in the 787 case a low voltage
has been chosen and there will be no need for such a thick insulation layer.
Moreover, during the DEAP project even more optimistic assumptions were
made regarding the weight per unit length of MgB2 cables reaching values even
close to 1kg/m (Wright et al., 2015). Nevertheless, for the purposes of this study
a more conservative estimation was chosen. It is important to note that in case
that these values could be reached, this will result in an important extra weight
benefit. Taking all these limitations under consideration it can be seen that there
is an almost 200kg (current aircraft) to almost 100kg (future conventional)
benefit on the cables’ weight based on the NASA predictions. On the other
hand, the DEAP estimates give a more significant weight reduction over both
conventional cases which could be highlighted even more if the optimistic
assumptions were used.
138
Figure 73 Electric loads and efficiencies diagram of the 787 electrical power
network in SEA case (DEAP estimates)
The weight of the required cooling system has been considered as a main
barrier for the superconducting cases. The results of this study however do not
exactly confirm this statement. It is indeed a heavy subsystem but in the
example of the existing 787 aircraft and based on the NASA efficiency figures
(Figure 74), the overall weight of the required cryo-coolers will be less than two
times that of the conventional cooling system. The difference becomes more
significant if the DEAP efficiency estimates are considered (Figure 73) and even
more enhanced when a comparison with the future conventional case is made.
It should be noted however that the most conservative estimates regarding the
operational temperature were chosen. An operational temperature of 20K was
chosen for the whole system. This might be the case at least for the electrical
machines which at this stage it seems inevitable to use MgB2 material for their
stator. However, the power electronics could be operated in higher-still
“cryogenic”-temperatures. In fact, as it was described in section 5.2.2, most of
the power converters might operate better in temperatures close to 100K. This
will significantly decrease the amount of required cooling power. Furthermore,
superconducting transmission lines using BSSCO as their main material with an
operating temperature around 90K might be used. With future improvements,
BSSCO cables could be the ideal option for airborne applications mainly due to
139
their increased critical temperature. The mass of the cryo-coolers, if these
higher operational temperatures for the power converters and the main cable
span are chosen, will be significantly lighter. In this study, both for reasons of
simplicity and conservatism, a uniform operational temperature of 20K was
selected.
Figure 74 Electric loads and efficiencies diagram of the 787 electrical power
network in SEA case (NASA estimates)
To sum up, it can be seen that SEA compared to the current MEA example of
787 aircraft would be a lighter and more efficient option both using the more
conservative estimates of DEAP program but also with the optimistic NASA
predictions. On the other hand, if these two superconducting versions of 787
are compared to the possible future version of a 787 type of aircraft there is no
weight benefit derived from the use of superconducting components but instead
around extra 50kg (NASA estimates) to 80kg (DEAP estimates) will be added to
the secondary power network of this aircraft.
In efficiency terms however there is still a clear benefit of using a SPN.
According to NASA efficiencies up to 99.7% can be reached in the electrical
power network of such an aircraft. This number goes down to 98.9% if the
140
DEAP estimates are used instead. Even this efficiency however is almost 2%
better than the one derived by using the optimistic future expectations for the
conventional room temperature components. Although this difference might
seem insignificant, it can result in excessive heat loads especially in the case
that the required electric load of these aircraft increases. A 90% efficiency,
which is currently assumed for the 787 aircraft in service, is a value that cannot
be considered competitive with all the other design options for the 2035
timeframe.
Apart from the efficiency gains there is another important advantage of using
superconducting components in the system under investigation. There are
reasons to believe that the implementation of the MEA approach to aircraft of
different sizes has been blocked by the fact that the scalability of the electrical
components of the network is not proportional to the one of the aircraft itself.
Power electronics and transmission line cables do not scale accordingly to the
aircraft size and this fact complicates the design optimisation of such an aircraft.
The next section (5.4) will investigate how the weight of the main components
will change depending on the electric power load demand. The cases of
increased electric power demand will correspond to either larger aircraft or/and
to further electrified versions of future civil MEA.
5.4 SEA Sensitivity/Scalability Studies
In this section the SEA study will be extended to different aircraft sizes. State of
the art aircraft will be used as references. The total required electric load for
each aircraft will be decided based on a factor that will determine the electric
power demand depending on the number of passengers that each aircraft can
carry. More specifically, the value of this factor is derived from the 787 model
where 242 passengers required 1MW of electric power. Hence, for each case
the total electric power demand was estimated using the following equation:
𝑇𝑜𝑡𝑎𝑙 𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐 𝐿𝑜𝑎𝑑 = 𝑃𝐴𝑋 ∗ 𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐 𝐿𝑜𝑎𝑑𝑓𝑎𝑐𝑡𝑜𝑟
(5-2)
141
Where PAX is the number of passenger of each aircraft and 𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐 𝐿𝑜𝑎𝑑𝑓𝑎𝑐𝑡𝑜𝑟
equals to 4.132 kW/passenger based on the 787 aircraft requirements.
This is just a simplistic method to estimate the potential electric load of different
aircraft assuming that similar secondary power network architectures will be
used in each case. This means that the same four different buses are assumed
to be part of the secondary power network of these aircraft. Although this might
not seem realistic since each aircraft design could be optimised differently it is
reasonable to make such an assumption in this early stage sensitivity study.
5.4.1 Reference Aircraft Description
Five different aircraft were chosen as representative examples of different
sizes/ranges commercial airplanes examples. In the near future Boeing is
looking to release updated versions of their 737 and 777 models (Scott, 2014).
The most recent Airbus aircraft are A350 and A380 models, whilst Bombardier
in 2008 put into production their regional commercial airplane CRJ-1000. In this
subsection a brief description of each reference aircraft will follow (www.airlines-
inform.com, 2012) :
Boeing 737
The Boeing 737 family is the most commercially successful family with more
than 4000 units sold. The latest model of this family is the 737-900. The
following table summarises its main characteristics:
142
Table 15 BOEING 737-900 Main characteristics
Variable Units Value
Range km 5080
Passengers - 189
Engines lb 2*27300
Maximum speed km/h 1000
Expected Electric
Load Demand
kW 780
Boeing 777
777 is a long range aircraft that entered production in 1995 and it flies to the
largest international airports. The following table summarises the main attributes
of the 777-300 which held its first flight in 1997:
Table 16 BOEING 777-300 Main characteristics
Variable Units Value
Range km 11000
Passengers - 550
Engines lb 2*115000
Maximum speed km/h 945
Expected Electric
Load Demand
kW 2275
Airbus A350
A350 is the newest aircraft in service, entering on 15 January 2015 with Qatar
Airways. It is a long range aircraft that was developed to succeed the A330 and
143
A340 and compete with Boeing’s 787 and 777 models. Its main characteristics
can be seen in Table 17:
Table 17 A350 Main characteristics
Variable Units Value
Range km 14800
Passengers - 475
Engines lb 2*95000
Maximum speed km/h 945
Expected Electric
Load Demand
kW 1965
Airbus A380
A380 is the largest passengers’ aircraft in the world that entered the commercial
service in 2007. A more detailed description of this model has already been
made (5.1.2) but the more important attributes in regards to this study can be
found in the table that follows:
Table 18 A380 Main characteristics
Variable Units Value
Range km 15000
Passengers - 700
Engines lb 4*70000
Maximum speed km/h 1070
Expected Electric
Load Demand
kW 2895
144
Bombardier CRJ-1000
As a final reference a Bombardier’s aircraft was chosen. CRJ-1000 is a regional
airliner and it will be used as an example of minimum electric power demand in
this study. Table 19 gives the main characteristics of such an aircraft:
Table 19 CRJ-1000 Main characteristics
The difficulties of designing an all new aircraft have pushed the biggest airliners
today to focus on updates on their existing products. Both Boeing and Airbus
have already announced their perspective passenger jets. The former is
planning to release 737Max and 777X as improved models of the already
existing family, whilst Airbus is launching their Airbus 330neo as the next
aircraft to be released after the recent A350 delivery to service (Shankland,
2014). Therefore, the choice of all the aforementioned reference aircraft models
was based on the future trends of the top aircraft makers as well as on the fact
that a wide range of aircraft sizes was necessary to be investigated.
5.4.2 MEA and SEA Simulink models overview
This sensitivity study was then extended for different aircraft sizes and electric
loads with the use of Simulink models for each case separately. Figure 75
demonstrates the models being used for the conventional MEA cases. There
are four main subsystems dedicated to the four components under investigation
Variable Units Value
Range km 2760
Passengers - 100
Engines lb 2*13630
Maximum speed km/h 880
Expected Electric
Load Demand
kW 415
145
(i.e. machines, power electronics, cables, and cooling system). The main inputs
and outputs of this model can be found on the following table.
Table 20 Main inputs and outputs of MEA’s power network Simulink model
Inputs Units Outputs Units
Total Electric Power Load 𝑊 Total VFSGs’ weight 𝑘𝑔
Number of Engines − Total VFSGs’ thermal load 𝑊
VFSG’s efficiency % Total P.E.s’ weight 𝑘𝑔
VFSG’s power density 𝑘𝑊/𝑘𝑔 Total P.E.s’ thermal load 𝑊
P.E.’s power factor − Cooling System’s weight 𝑘𝑔
P.E.’s power density 𝑘𝑊/𝑘𝑔 Cables’ weight 𝑘𝑔
P.E.’S efficiency % Total system’s weight 𝑘𝑔
Cooling Weight Factor − Total system’s thermal load 𝑊
Cable Length 𝑚
Nominal System’s Voltage 𝑉
Maximum Current Capability 𝐴/𝑚𝑚2
The total electric power load and the number of engines in each aircraft can be
found in 5.4.1. Depending on the number of engines, the number of VFSGs
(two per engine) and consequently the power rating of each machine could be
estimated. Using the assumed values of VFSGs’ power densities and
efficiencies the weight and total thermal load of the machines could be
estimated. In this study, the total electric power demand was distributed to the
several buses in accordance to the current conventional 787 case. Hence,
48.9% of the total load was delivered in the +/- 270VDC line, 20.22% is
transmitted to the 115VAC bus bar, only 4.28% is used to satisfy the +/- 28VDC
loads, whilst the rest 26.6% of the total electric power available is used to power
the remaining 230VAC secondary loads. These power factors were used as
inputs to the power electronics subsystem in order the required power rating of
146
each converter to be predicted. As it was stated in previous sections, a cooling
weight factor of 0.3 was assumed in order to estimate the conventional cooling
system’s weight. Furthermore, the main transmission line’s total length for each
aircraft (5.4.1) combined with the system’s nominal voltage (230VAC), and the
maximum current capability of copper wires (4 𝐴/𝑚𝑚2) were the transmission
lines subsystem inputs. Finally, combining the outputs of each subsystem, the
MEA’s power system total weight and thermal load could be estimated for both
conventional cases (current and future technology).
Figure 75 MEA’s Electric Power Network Simulink Model
For the superconducting cases a more complicated power network model has
been developed (Figure 76). The superconducting electrical machine and the
cryo-cooler models have been extensively analysed in 4.2 and Appendix A.2
respectively. The power electronics’ subsystem is the same as in the
conventional cases, whilst the main transmission lines’ weight and total thermal
load is being calculated by using simple weight and losses per unit length
values (presented in 5.3.1). The only difference between the models of the
superconducting cases is in the way the machines’ weight is calculated. In the
NASA cases a simple torque density value is being used to estimate the
machine’s weight instead of the complicated fully superconducting machine
models used in the DEAP case. The majority of inputs and outputs of SEA
147
power network model are the same as the ones in Table 20. However, there are
a significant number of additional inputs due to the complex superconducting
VFSGs and Cryo-cooler models. The inputs of these subsystems can be found
in the respective chapters where they were presented (i.e. 4.2.4 and Appendix
A.2).
Figure 76 SEA’s Electric Power Network Simulink Model (Superconducting DEAP
case)
5.4.3 Weight Trends in reference aircraft
The next step for this study will be the comparison of the weight of the various
components for each reference aircraft using the same four different versions of
secondary power networks used in 5.3 and the Simulink models presented in
the previous section. The various components will be investigated separately,
whilst a total system’s weight comparison will follow.
Electrical machines
The weight of the VFSGs in each case was calculated by using the majority of
the assumptions presented in section 5.3.1. Depending on the electric power
requirements of each aircraft the nominal power rating of the machines in each
case was calculated. In the superconducting cases the machines were
adequately oversized in order to deal with the cryo-coolers power demand.
148
Moreover, two generators per engine were assumed in every reference aircraft.
In the current technology case a torque density of 3.65 N*m/kg was assumed
based on the VFSGs that are used in the 787 aircraft. In order to simplify this
study all the generators were assumed to have the same variable frequency as
the initial reference aircraft (i.e. 360-800 Hz). The following figure presents the
total weight of the electrical machines for each reference aircraft in all four
different versions.
Figure 77 VFSGs’ weight for each reference aircraft in all four different versions
As it was expected the total weight of the electrical machines increases as the
electrical power demand rises. The benefit of using superconducting machines
is particularly highlighted for larger aircraft-such as the A380, B777, and A350-
where the machines in the superconducting versions are anticipated to be over
an order of magnitude lighter (with NASA estimates) than the ones currently
used in 787. Three times lighter machines (NASA values) are expected in
comparison to the optimistic estimates for the future technology in the long
range type of aircraft. The moderate estimates of the DEAP program slightly
decrease the weight benefit compared to the current technology figures;
however there is still an important difference in these two versions. On the other
hand, there is hardly any benefit compared to the future trends especially for the
short range aircraft (i.e. CRJ-1000 and B737).
0
200
400
600
800
1000
1200
Current Technology Future Technology Superconducting(DEAP)
Superconducting(NASA)
Wei
ght
(kg)
CRJ-1000 B737 B787 A350 B777 A380
149
To sum up, the advantage of using superconducting machines in terms of
weight savings in undeniable compared to the current technology being used in
such an advanced aircraft as “The Dreamliner” example. However, if the
relatively optimistic predictions for future technology of the electrical machines
will be verified then the resulted benefit will be significantly limited. It is
important to note that for the future technology a power density factor is used
due to available data in the literature. However, this is not a common method for
calculating the weight of the electrical machines, since the torque is what
typically sizes these machines. In this study the frequency-and hence the
rotational speed-of the machines was held constant in each case so that the
torque and power density will change accordingly.
Power Electronics
The same assumptions as the ones summarised in Table 13 were used for the
weight calculation of the main power converters necessary in each reference
aircraft. As it was previously stated the same architecture was assumed for
each aircraft although this does not correspond to reality at the moment.
However, it seems reasonable that if a more electric approach will be followed
in all the aircraft under investigation a similar electrical power network to the
existing one of 787 will be most probably used. In any case, the majority of the
electric load produced by the VFSGs will have to be converted to different
voltage levels to be useable for the secondary loads of the aircraft. Figure 78
demonstrates the Simulink subsystem for the weight and thermal load
calculation of the power electronics in the MEA and SEA concepts. The inputs
of this subsystem can be found in the overview Simulink models previously
presented in 5.4.2 and differ based on the total electric load of the aircraft, as
well as its nature (conventional or superconducting).
150
Figure 78 Power Electronics’ Simulink Model
The results of the power electronics’ weight study are presented in the following
figure:
Figure 79 Power Electronics’ weight for each reference aircraft in all four
different versions
The outcomes of this study resulted in similar trends as the electrical machines’
study. There is a clear benefit of using cryo-cooled power electronics in
0
100
200
300
400
500
600
Current Technology Future Technology Superconducting(DEAP)
Superconducting(NASA)
Wei
ght
(kg)
CRJ-1000 B737 B787 A350 B777 A380
151
comparison to the current technology that results in almost seven times lighter
converters. Since this study was practically a power density examination it is
easy to make similar conclusions for each case. The DEAP estimates and the
conventional future technology expectations differ only by 5kW/kg per
converter, a difference that could prove to be significant only in high power
applications. Over two times lighter equipment is expected with NASA estimates
compare to the future technology predictions.
To conclude, power electronics weight seems able to block the feasibility of
MEA approach particularly in long range aircraft where they could add close to
half a tonne in the system (A380 case). However, if the future expectations for
the conventionally cooled power electronics could be met then the
superconducting cases will not result in significantly lighter equipment.
Cables
First of all, the length of the main transmission lines needs to be estimated in
each reference aircraft. In the case of Boeing 787 a length of 75 m for the main
transmission lines was chosen based on the aircraft’s length and the location of
the engines on the wing. Since there is not a clear idea of the exact electrical
architecture in each aircraft this method just gives an approximation of the
required length. Following a similar strategy for each reference airplane the
required length of cables was calculated in each case. Equation (5-1) was then
used to calculate the weight per meter of the conventional copper cables (with a
10% improvement for the future technology). In the superconducting versions
the weight per meter of the power cables is assumed constant in all cases. The
required thermal insulation in this type of transmission line is expected to be the
main weight factor in the low power applications, whilst the increased current
density capabilities of superconducting wires will allow them to keep the size of
their superconductor relatively constant to any power changes. When the
normal currents are relatively low (i.e. CRJ-1000, B737 cases) using thin
superconducting wires will make the connections and mechanical support a
challenging task (Malkin and Pagonis, 2013). Table 21 summarises the weight
152
density of the main span cables in each reference aircraft for all the different
versions under investigation.
Table 21 Weight per meter (kg/m) of the main transmission lines for each
reference aircraft in all four different versions
Aircraft Length
(m)
787
Current
787
Future
Superconducting
Case (DEAP)
Superconducting
Case (NASA)
CRJ-1000 65 4.8 4.3 5 8
B737 70 8.5 7.7 5 8
B787 75 10.5 9.5 5 8
A350 80 20.5 18.5 5 8
B777 90 24.4 21.9 5 8
A380 120 29.4 26.5 5 8
The total weight of these cables can be seen in Figure 80 where it shows that
the benefit of having superconducting cables can be capitalised for aircraft
larger than the Boeing 737 model. This weight benefit is extremely highlighted
in examples such as the Boeing 777 and Airbus A380 where both the aircraft
length and the aircraft electric power demand have a detrimental effect on the
weight of conventional copper power cables. This weight benefit can reach
values over a tonne in the case of A380.
153
Figure 80 Cables’ weight for each reference aircraft in all four different versions
Finally, it is important to note that in both superconducting versions losses in the
order of 5 W/m of cable length were assumed (Brown, 2011). This parameter is
important for the calculation of the thermal load that the superconducting
configurations produce. The latter will be the predominant factor on the weight
estimation of the required cooling system.
Cooling system
In this section the weight of the cooling system in each case will be estimated.
In the conventional configurations, where room temperature equipment is used,
the cooling system was assumed to weigh 30% of the overall weight of the
cooled components. This approximation was used since no relative information
was found in the literature and it is based on an expert’s opinion (i.e. Stephen
Harrison). In regards to the superconducting versions of each reference aircraft
the two-stage reverse Brayton cryo-coolers-presented in 5.3.1 and fully
described in A.2- were used. NASA has made some approximations regarding
the expected cryo-cooler weight assuming a power density around 5 lb/input-hp.
However, the cryo-cooler models developed during the DEAP project can be
considered as more reliable than a simple power density assumption. Hence,
these models were used for both superconducting cases giving a more constant
0
500
1000
1500
2000
2500
3000
3500
4000
Current Technology Future Technology Superconducting(DEAP)
Superconducting(NASA)
Wei
ght
(kg)
CRJ-1000 B737 B787 A350 B777 A380
154
representation of the cooling system in these cases. Aircraft estimations based
on the NASA assumptions will nonetheless benefit from the increased
components’ efficiency which will result in lower thermal loads.
Figure 81 Cooling system’s weight for each reference aircraft in all four different
versions
As it was expected in the superconducting versions the weight of the cooling
system is significantly higher in most of the cases. However, it should be noted
that as the electric load increases the difference between conventional and
superconducting cooling system is dramatically decreasing. For example, in the
A380 type of aircraft the cryo-coolers’ weight based on the DEAP efficiency
assumption will weigh less than two times the conventional cooling system
weight, whilst if the NASA efficiency figures are used there is only a 30 kg
penalty in the superconducting modification of the aircraft. Clearly, this changes
if the future technology predictions are used as a comparison to the
superconducting models. In that case, the cryo-coolers’ weight can be up to ten
times heavier if the NASA estimates are used or even up to 17 times heavier by
using the DEAP efficiency predictions (CRJ-1000 model).
It is becoming clear that the weight of the required cryo-coolers does not
increase linearly to the overall electric load of the aircraft. In the contrary, the
0
100
200
300
400
500
600
700
800
900
Current Technology Future Technology Superconducting(DEAP)
Superconducting(NASA)
Wei
ght
(kg)
CRJ-1000 B737 B787 A350 B777 A380
155
longest the range of the aircraft the less effect on the overall weight of the
system the cryo-coolers weight will have. This result was somehow anticipated
but the extent of this effect was highlighted via this study.
5.4.4 Final Remarks
Figure 82 summarises the overall estimated weight of the secondary power
network of each reference aircraft based on current and future technology
component density estimations as well as superconducting component weight
predictions made both by the DEAP project team and NASA.
Figure 82 Electrical Power Network total weight for each reference aircraft in all
four different versions
The first thing to notice is the different scalability ratios between conventional
and superconducting versions. Although in the short/medium range aircraft
models (i.e. CRJ-1000, B737, and B787) the overall weight of the conventional
electric power network is comparable or even lighter than the superconducting
equivalent, in the longer range aircraft examples (i.e. A350, B777, and A380)
the SEA concept becomes a very attractive option in terms of the overall
system’s weight and efficiency even if the aggressive density targets for the
conventional equipment could be reached. If instead of specific aircraft models
the electric load requirement was used as a reference, it seems like the 1.5 MW
0
1000
2000
3000
4000
5000
6000
Current Technology Future Technology Superconducting(DEAP)
Superconducting(NASA)
Wei
ght
(kg)
CRJ-1000 B737 B787 A350 B777 A380
156
is the limit where the superconducting case gives an overall weight benefit
compared to the future conventional technology case (Figure 83). Hence, as the
electrification of future aircraft is expecting to rise, even shorter range aircraft
could benefit by the use of superconducting components. Figure 83
demonstrates the weight trendline compared to the electric load requirements of
an aircraft for all four cases. It is clear that as the electric load demand rises the
superconducting cases become more attractive options in terms of weight.
Figure 83 Electrical Power Network total weight for different electric load
requirements
This weight benefit derives from the non-linear way in which cryo-cooler weight
is increasing with the cooling power demand. Another thing to note is that if
higher operational temperatures were chosen for components such as the
power electronics and the main transmission cables the cooling power demand
and consequently the cryo-cooler weight would have been significantly lower.
Furthermore, the main transmission lines start to take advantage of the
attractive characteristics of superconducting wires mainly as the nominal current
of the system significantly increases. As it was described in Chapter 3 the way
superconducting networks are designed is relatively different than the
procedure in conventional power networks. High normal currents are actually
preferred in these cases due to the incredibly high current capability of
0
1000
2000
3000
4000
5000
6000
0 500 1000 1500 2000 2500 3000 3500
Wei
ght
(kg)
Electric Power Demand (kW)
Current Technology Future Technology
Superconducting (DEAP) Superconducting (NASA)
157
superconductors. The acceptable range in voltage levels in an airborne
application is not wide enough to allow lower normal currents in the
conventional configurations and this fact has a detrimental effect on the weight
and size of the cables in long range aircraft where a MEA approach is followed.
Finally, whilst in a MEA the use of a SPN could be considered as optional at
least for shorter range aircraft, in the TeDP or HEDP concepts their use seems
inevitable. There is at least an order of magnitude difference in the electric
power requirements between the two approaches and the potential weight and
efficiency gains of a SPN make their use necessary in the case of HEDP.
Hence, if SPNs were adopted by MEA this will make the eventual transfer to
hybrid/electric more progressive (Malkin and Pagonis, 2015a).
5.5 Key Study Limitations
This study could be considered as a preliminary feasibility study of the SEA
approach in existing and future aircraft. The SEA concept includes a number of
components that have not been built yet and hence a number of assumptions
were necessary. Therefore, there are a number of factors that limit the accuracy
of this investigation and are presented in this subchapter.
The majority of the superconducting components are still in the early
stages of development with low technology readiness levels (TRL) of
0-2. The same goes for the future technology estimates where
aggressive power density values were assumed. Although these
assumptions are adequate for preliminary weight studies, a lot of work
needs to be done so that the technology could be considered mature
enough to be implemented in an aerospace application. In regards to the
superconducting electrical machines and cryo-coolers models the key
limitations have been described in chapter 4.4 and A.2 respectively.
An operational temperature of 20K for the whole system in the
superconducting cases was assumed. This temperature was chosen for
two main reasons: a) to explore the worst case scenario in terms of
cooling demand and b) to simplify the architecture of the cooling system.
Having different component operational temperatures will have resulted
158
in separate branches of cryo-coolers adding complexity and possibly
weight in the electrical power network. However, it is fair to assume that
higher operational temperatures will be achievable both for the power
electronics and for the main transmission lines resulting in a significant
drop on the cooling power demand. The fully superconducting machines
on the other hand might need to operate on the 20-25 K range due to the
AC losses of the stator that can be acceptable only in the case of MgB2
material.
A final remark about the cooling system is the possibility of using a
liquid cryogenic fluid and a heat sink instead of the bulky cryo-coolers
being investigated in this study. The main benefit of such a cooling
method could be the use of a coolant fluid from which the boil-off gas can
be also used as a low emissions fuel (Malkin and Pagonis, 2015a).
Especially, if an operating temperature of 111K was chosen for the
power electronics, LNG could be used reducing the overall costs
significantly. However, the investigation of the optimal cooling system
was not in the scope of this research study. The optimisation of this
system could enhance even more the feasibility of the SEA concept.
The main transmission line’s weight -especially for the
superconducting cases- was based on generic assumptions of low
fidelity. In the superconducting versions, a constant weight per meter
value was chosen, neglecting any effect the different power, voltage and
current levels might have. Notwithstanding these remarks, it should be
noted that the reference superconducting cables were characterised
either by their high voltage (Xi et al., 2006) or their high power (Wright et
al., 2015) levels; hence, they could be considered as moderate
assumptions.
There are also other components of the aircraft’s secondary power
network that have not been included in this weight sensitivity study.
Switching and protection devices for example could add significant
weight in the whole system both in the conventional and in the
superconducting cases. Equipment devices such as circuit breakers,
159
SFCLs, and solid-state switches will be necessary in these configurations
and an extended study of the ideal protection system could modify the
optimal electrical system architecture for different aircraft. It is not clear
yet if there will be a significant weight and efficiency difference between
the conventional and the superconducting configurations.
Last but not least, the electrical system architecture of each aircraft
could be optimised differently. The Boeing 787’s electrical power
network was used as a baseline since it was the only existing
architecture implementing the MEA concept. Nevertheless, as the
electric load demand increases alternative design routes might be
followed. In regards to voltage levels as it was stated in Chapter 3,
Paschen Law’s limits the maximum voltage level acceptable for an
airborne application to approximately 327 V (Armstrong et al., 2012).
MEA and SEA will most probably follow different design approaches but
the most favourable one in each case will be decided after extensive
sensitivities studies.
160
6 Novel Flight Cycles for Hybrid/Electric Aircraft Using
Energy Storage
Hybrid/electric approaches have attracted the interest of many industries (i.e.
marine, automotive etc.) mainly due to the benefits derived from the flexibility
they offer in the operating cycles. Here in aerospace field the motivation
towards HEDP approach is forced by the improvements in propulsive efficiency
and aerodynamics. In this chapter we will look at aircraft operational cycles to
investigate if any additional improvements can also be attained.
The previous chapters had already shown some of the potential benefits that a
TeDP configuration could offer. However, it was apparent that extra benefits
could be obtained if an overall novel optimised system’s approach is followed.
The hybrid/electric nature of the proposed configuration could free the
propulsion system of this type of aircraft from the restrictions that conventional
configurations are facing. The main approach is based on the fact that each
propulsive unit could be optimised for a specific function (propulsive or not)
increasing its efficiency throughout the flight cycle. The optimisation of an
electric power network increases the flexibility of the whole system and this is
one of the main advantages of these configurations that up until now have not
been fully explored.
Energy storage could play an important role on these novel designs adding
even more flexibility to the whole network. These devices could be used either
as a short term power unit, as a boost power source or as a main prime mover
depending on the range and power requirements of the aircraft. Batteries,
supercapacitors and Superconducting Magnetic Energy Storage (SMES) are
some of the energy storage options for the flexible integrated power system
under investigation.
Several architecture proposals and novel flight cycles will be explored in this
chapter. The DEAP aircraft will be used as a reference aircraft and the weight
estimation of the main components of these novel configuration proposals will
be calculated so that the feasibility of these designs to be determined. After
161
that, the study will be extended to different aircraft sizes from short to long
range cases.
6.1 Energy Storage
Energy storage mechanisms are currently been used in numerous applications
including the main transportation industries (i.e. aerospace, automotive, marine
etc.). More electric or all electric vehicles are continuously attracting the interest
of these industries that are eager to create more efficient and environmentally
friendly vehicles. Energy storage devices are present in all these more electric
configurations with different requirements each time depending on the
application. Batteries is clearly the most mature and well-established technology
of energy storage. However, other storage mechanisms such as the
supercapacitors seem to improve rapidly creating competitive to battery
products especially for specific applications where quick and short term power
demands are necessary. SMES could also play an important role in a concept
such as the TeDP, since they can be integrated in the already existing
superconducting network increasing remarkably their actual power and energy
densities.
6.1.1 Batteries
State of the art
A study about battery technology is highly dependent on the application.
Batteries can be divided into two main types depending on their charging
capability: primary and secondary (or else rechargeable). In this study only the
latter type will be investigated. In this category, at present, there are four main
types that have been broadly used in the industry:
Lead-acid
Nickel-Cadmium
Ni-Metal Hydride
Lithium-Ion
Since the study is made for airborne applications the energy density of the
battery becomes the crucial feature. In addition, safety, life cycle, and reliability
162
are also important factors. Figure 84 presents the theoretical and the current
practical specific energy values of the four aforementioned types of battery.
Figure 84 Current Specific Energy values of different battery types
*Note that multiple Li-ion technologies are currently commercially used and the values given in the figure
are just an average of the best cases
A more detailed table which includes several important factors for a vast variety
of battery types can be found next:
Table 22 Comparison of different types of battery currently in use
(www.batteryspace.com, 2015)
0
50
100
150
200
250
300
350
400
450
Wat
t-h
rs/k
g
Theoretical Value
Practical Value
163
The table above is useful for making some important conclusions. Looking
exclusively at the energy density values lithium cobalt oxide (LiCoO2) seems the
most attractive option. However, this type of battery is characterised by its high
cost and most importantly as being unsafe for high power applications such as
the aerospace. It is normally used in portable applications, being the most
common option in mobiles, laptops and cameras. A far safer option and with
the extra benefit of long cycle life and extremely low cost is the lithium iron
phosphate (LiFePO4) battery. However, its energy density is limited to around
120 Wh/kg lower that other lithium-ion types.
Focusing on similar to aerospace applications could be more useful to
understand the current technology trends for batteries. Electric vehicles seem
the closest application to aircraft propulsion with similar priorities in the choice
of energy storage. Toyota Prius 04 uses a prismatic NiMH battery with specific
energy density of 46 Wh/kg (www.eaa-phev.org, 2015). The Chevy Volt on the
other hand uses a Li-ion battery pack with specific energy of 53 Wh/kg (Murphy,
2012). Finally, Nissan Leaf uses a laminated lithium ion pack with energy
density around 140 Wh/kg (wikipedia.com, 2015) which seems to be the most
reliable state of the art example of battery that could also be used in an
aerospace application. New Li-ion cells for automotive applications are under
development with the examples of Saft VL45E and VL41M (with energy
densities of 160 and 146 Wh/kg respectively) to present the most attractive
characteristics (Rosenkranz, Kohler and Liska, 2007).
In conclusion, the decision about which battery technology to use in a TeDP
configuration is more complex that might have been believed. Although lower
weight, and as a consequence higher energy density, is the number one priority
other factors such as safety, life expectancy, cost and power delivery cannot be
neglected. The author believes that the most appropriate method of estimating
current energy densities of batteries is by comparing the battery types being
used in similar applications. Thus, electrical vehicles which have similar
priorities -such as high energy density, safety, life, and high power density-were
chosen as a reference. As a result, a specific energy around 200 Wh/kg and a
164
specific power performance up to 4000 W/kg can be considered as the state of
the art values for batteries (Jow et al., 2014). These two values will later been
used as the state of the art Li-ion parameters for the novel flight cycle study.
Figure 85 summarises the specific energy values of the different battery
technologies currently been used in the industry. Lithium Nickel Cobalt
Aluminium Oxide battery is the clear winner in regards to energy density storing
more capacity than any other technology (close to 260 Wh/kg), however it
suffers in terms of power density and thermal stability. Hence, the values being
used in this study as the state of the art density limits are closer to the ones
presented in Nickel-manganece-cobalt (NMC) and Lithium Cobalt technologies.
Figure 85 Typical specific energy values for different battery technologies
(batteryuniversity.com, 2015)
Future trends
Lithium-ion batteries have not yet reached their optimum performance and their
technology is continuously improving. Since the anode in lithium-ion type of
battery has been optimised, then batteries are cathode limited devices and
further developments in the cathode materials could lead to better battery
performance. However, there is a practical limit of lithium-ion battery capability
which even if it is attained it would still not provide the required energy density
165
for aerospace applications. There have been research groups aiming on
providing rechargeable Li-ion batteries with energy densities around 400 Wh/kg
and power densities up to 16000 W/kg (Jow et al., 2014).
Another type of battery which has gathered a lot of interest lately and it is
believed by many researchers to be the future of battery technology is the so
called lithium-air battery. The main difference between a lithium-ion and a
lithium-air battery is that the cathode is replaced by air making the latter type
significantly more lightweight and with greater energy capabilities (Ayre, 2014).
Lithium air batteries have a theoretical limit of around 12 kWh/kg without the
oxygen mass, a value comparable to the one of gasoline (~13 kWh/kg)
(Imanishi and Yamamoto, 2014). Li/Air technology is nearing commercialization
and has already achieved specific energies in excess of 700 Wh/kg (PolyPlus,
2009). However, there are still many problems that need to be addressed such
as the low discharge rate, poor life cycles, and low efficiency (Shen et al.,
2013). Power density of this type of batteries is also relatively low. Depending
on the degree of hybridization a target for a 140 to 1400 W/kg power density in
battery level has been set (Christensen et al., 2012).
Finally, one of the most promising high specific energy battery types is the
lithium-sulphur (Li-S). Li-S batteries present a theoretical specific energy five
times greater than the Li-ion technology (i.e. 2500 Wh/kg) (Shuli and Zhan,
2015). They hold the record in the highest specific energy density being
achieved to date by rechargeable batteries in an actual application (350 Wh/kg
for the Qinetiq’s Zephur UAV) (Millikin, 2010). Their relatively low cost makes
them even more attractive for potential extensive use. One of the main
drawbacks however is their low cyclability. Many studies and researchers have
been focused on increasing the life cycles of this type of batteries and many
labs have claimed that a 500 Wh/kg commercialised Li-S battery will soon be
available (Van Noorden, 2014), (Dodson, 2013). Extremely high power densities
in the range of 11000 W/kg after 100 cycles have been claimed for an all solid
state Li-S battery (Nagata and Chikusa, 2014). The benefits and drawbacks of
this technology can be seen in Figure 86. Specific power and energy are
166
extremely high, whilst the cycle life of these batteries is the field that urgently
needs improvement.
Figure 86 Status of Li-S batteries compared to the United States Advanced
Battery Consortium (USABC) baseline standards (Mikhaylik et al., 2015)
Battery technology keeps improving throughout the years bringing the future of
all or more electric applications closer to reality. In this section the most
promising technologies have been briefly described. Table 23 gives a summary
of the specific energy and power of these promising technologies. These values
will later be used in the sizing models developed in Simulink which will assess
the feasibility for novel flight cycles approach for the future hybrid/electric
aircraft. The table also includes an approximation of the expected cyclability of
the technologies under investigation. The Li-ion values mainly depend on the
specific technology being chosen; hence the wide range. In any case, the life
cycles expectance of Li-ion batteries is significantly higher than the other two
technologies. There is a general interest in improving Li-air cyclability figures
(Roveglia, 2015), whilst Li-sulphur cycle life is as Figure 86 shows the main
drawback of this technology.
167
Table 23 Battery technology summary and sizing parameters
Type Specific Energy
(Wh/kg)
Specific Power
(W/kg)
Life Cycles
Li-ion
(current)
200 4000 400-1200
Li-ion
(future)
400 8000 400-1200
Li-air 750 700 50-150
Li-sulphur 500 10000 ~100
6.1.2 Supercapacitors
Another mean of energy storage that is starting to attract more interest in the
industry is the capacitors; and more specifically the supercapacitors (also
known as ultracapacitors). Unlike batteries, who store energy in chemical
reactions, capacitors store energy in an electric field which is created between
two oppositely charged particles when they are separated by a dielectric.
Supercapacitors use a different storage mechanism to traditional electrostatic
capacitors, but behave in a similar way due to the way they store the energy. In
the vast majority of supercapacitor applications today (almost 95%) Electric
Double Layer Capacitors (EDLCs) are used with carbon as the active electrode
material (Simon and Gogotsi, 2010). In order to increase the energy stored in
these devices it is essential to increase the surface area. This surface storage
mechanism is partially one of the reasons for the relatively low energy density of
supercapacitors (typically around 5Wh/kg) (Simon and Gogotsi, 2008). Other
reasons are the limited operating voltage range, the required thickness of the
separator, as well as some practical limits such as packaging and internal
losses (Edwards, 2011).
The main attractive attribute of supercapacitors is their ability to deliver all their
stored energy in a really short time (around 5 sec). Moreover, they are capable
168
of withstanding thousands of cycles; even in the range of 500.000 cycles with
less than 20% capacitance decrease (Mouser Electronics, 2015).
In airborne applications supercapacitors have been used in Airbus 380 for the
emergency door opening. On the other hand, in automotive world the general
trend is the hybrid/electric vehicles with a combination of EDLCs and batteries
for fast acceleration and breaking energy recovery which leads to an increase of
the battery life expectance. This might also be the future in airborne
applications, where a hybrid system with batteries and supercapacitors could
become the most beneficial configuration.
Li-based hybrid systems with nanostructured Li4Ti5Oi2/AC formation
(introduced by (Amatucci, Badway and DuPasquier, 2000)) were the first
ultracapacitors which reached energy densities of 10 Wh/kg and high power
densities. Since then, several companies and institutions have been studied
the lithium based capacitors option. JM energy Corp. started the mass
production of Lithium-ion capacitors in 2009 with gravimetric densities of 12-14
Wh/kg (Millikin, 2009). ACT (Advanced Capacitors Technology) also claimed to
have built LIC (Lithium-ion Capacitors) such as the so-called Premlis which
uses for the cathode a nanoporous carbon material and for the anode graphitic
carbon material doped with lithium ions. This capacitor device doubles the
energy density of the company’s existing products (Hampton, 2013). Premlis
5000 was initially developed in Bhutan for LED light applications that were used
in areas without electrical supply in the city. The energy storage per unit volume
of this LIC was 25 Wh/kg (Montgomery, 2012). This type of capacitors
combines the relatively high energy of Li-ion batteries and the high power of
EDLCs. Next figure partially presents the aforementioned conclusions:
169
Figure 87 Energy and power density of different energy storage options
(Hampton, 2013)
Another type of supercapacitors that have attracted a lot of interest recently is
the EDLCs that use carbon nanotubes (CNTs). Several studies have been
carried out to explore the potentials of the CNTs with extended efforts to
improve the specific surface area and/or the operational voltage range. For the
former a controlled oxidation of single walled CNTs had led to a high energy
density of 24.7 Wh/kg, while an operational voltage of up to 4V has been
achieved in supercapacitors with high purity CNTs in conventional organic
solvents leading to an energy density of around 94 Wh/kg (Kim, Chung and
Kim, 2012). However, it must be pointed out that the last results were calculated
possibly without taking into account the ‘dead components’ of the final device.
But even if we assume a 50% decrease due to internal losses and packaging a
supercapacitor of deliverable stored energy around 50 Wh/kg can be feasible in
the short term. In the aforementioned CNTs, there is also enough room for
improvement for their power densities which are relatively low for
ultracapacitors’ standards.
To sum up, it is really difficult to predict the future improvements in
supercapacitors technology. There are several research studies focusing on
different materials that have shown promising results. Electrode materials such
as activated carbon, carbon fibres, carbon aerogel, CNTs, and graphene are all
170
being currently explored and showing potentials of reaching competitive specific
energy values. However, it seems unlikely that they will ever approach energy
densities similar to the ones expected from the future battery technologies. On
the other hand, the impact of the future improvements is less predictable than in
the batteries case; hence supercapacitors should not be ruled out. Especially
due to their extremely high power densities they could be the ideal candidates
for several applications where high power is required for a short period of time.
In this study, a specific energy of 50 Wh/kg and a specific power of 15 kW/kg
were considered as realistic estimates for the 2035 timeframe.
6.1.3 Superconducting Magnetic Energy Storage (SMES)
The superconducting nature of the networks under investigation in this research
study makes the use of SMES a viable option for an aerospace application. This
system stores energy in the magnetic field created by the flow of direct current
in a superconducting coil (Sutanto and Cheng, 2009). SMES does not include
any energy conversion (pure electrical conversion only) resulting in fast
response times. Their efficiencies are relatively high and their capability of
unlimited discharges and recharges give them an extra advantage over
batteries. Moreover, they present a good balance between power and energy
density which could be important for an aerospace application. The main
advantage however, seems to be the capability of discharging large amounts of
power for a small period of time and unlimited times (Yuan et al., 2010).
Figure 88 Schematic of a SMES device (Molina, 2010)
171
The main disadvantage of these devices is their cost, as the materials being
used are normally too expensive to be considered in large applications.
Furthermore, the refrigeration system also normally poses an obstacle in most
applications. However, in a TeDP configuration a cooling system will already be
present; hence, the cooling power requirements of a SMES are not expected to
affect the feasibility of this type of storage. Finally, their mechanical instability
and their high self-discharge ratio for long periods are also significant concerns
(Yuan, 2011).
SMES systems are typically used to improve the network’s stability and power
quality. In our network, SMES could work as a supplementary source of energy
storage. More specifically, in a power failure their fast response allows them to
provide electrical power in the very few first seconds while other types of energy
storage could supply power later on.
6.2 Novel Hybrid Configurations and Flight Cycles
One of the major benefits of the Turbo-electric Distributed Propulsion (TeDP)
concept is the flexibility that offers to the whole system design. There are many
design options regarding the complete flight cycle of this new aircraft that have
not been investigated yet. Energy storage could play an important role on future
aircraft designs, especially if the technologies described earlier reach their full
potential. In a hybrid/electric configuration significant electrical power is being
used in the distributed propulsion system. Typically, this required power is
produced by the gas-turbine alternators which produce electrical power to
satisfy the demands of the whole network at all times. An aircraft mission profile
consists of five main flight phases: taxi, take-off, cruise, descent, and landing.
By looking on the power demand during these phases someone will notice a
significant peak during the take-off phase. Obviously, this fact puts constraints
on the engines’ design which are rated to satisfy this peak power whilst for most
of the mission are working at the half of their potential or less. The possibility of
using alternative power sources such as energy storage either to reduce the
power peaks of the engines or to optimise all the prime movers involved for
specific flight phases will be the main target of this chapter.
172
6.2.1 Baseline Aircraft and Mission Profile
For the purposes of this study and for reasons of consistency the “DEAP
aircraft” was used as a reference. This is a short/medium range aircraft with
capacity of around 100 passengers. A brief description of the DEAP project’s
aircraft has been presented in 1.3 and in 2.1.5, whilst its main characteristics
are summarised in Table 24. Note that the thrust requirements differ depending
on the configuration being chosen but this value is used to give an
approximation for the power requirements of this aircraft.
Table 24 Main characteristics of DEAP Aircraft
Characteristic Value Units
Mach Number 0.75 -
PAX 100 passengers
Range 2000 Nm
TOC 34000 Ft
Thrust
Requirements
~25 kN
The mission profile of the DEAP aircraft can be seen in Figure 89. The red
circles indicate the three main sensitive areas of the mission which need
optimisation during the design process of the GTs in conventional
configurations. The most crucial targets of the GTs are:
To satisfy the peak power demand of the mission (i.e. EOR-one engine
out safety case).
To satisfy the highest corrected flow at inlet to the compression system
(i.e. TOC).
To be as efficient as possible during the longest phase of the mission
(i.e. the cruise).
173
Figure 89 Mission profile of the DEAP aircraft
The former of the above three crucial design “hot spots” is an important criterion
as it effectively sizes the power system of the aircraft. In the proposed
configurations the system could be designed in the EOR-one engine out safety
case to land safely powered by one or two GTs combined with the use of E.S
subsystems. This eventually could reduce the size of the GTs allowing more
optimal GT design points that would probably increase the efficiency of the
engines (effect on the rest of “hot spots”). To cope with several design
challenges that may occur, the electrical system should be designed in a way
that a fully symmetrical thrust can be produced during the one engine out safety
case. Furthermore, the overrating of electrical machines could give significant
design benefits. More specifically, the electrical machines could withstand
excessive power demands for short periods of time, such as the safety case or
even during the landing phase. This will ease the design of such a system,
while following a similar strategy with the E.S. devices (i.e. higher power
discharge during the short climb phase) could potentially reduce the size and
weight of these components.
174
6.2.2 Overview of the Modelling Approach
This thesis’ section aims to investigate the feasibility in terms of weight of using
energy storage mechanisms throughout the flight cycle of a HEDP type of
aircraft. Other benefits and challenges will also be explored but a number of
Simulink models were developed to estimate the weight of the main
components of a hybrid configuration. The weight of components such as the
GTs, electrical machines, cryo-coolers, and energy storage devices will be
estimated for a number of different proposed configurations. Most of the models
were developed during the DEAP project, although a wider feasibility study in
terms of energy storage use has been investigated in this research study.
Although the DEAP project assumed geared turbofans as the main engines for
their proposed architectures, in this study the GTs are assumed to be
turboshaft, an assumption that was also used during the NASA N3-X studies. It
seems reasonable that since the main role of the gas turbines in a
hybrid/electric configuration will be the production of power-and not thrust-
turboshaft engines will be the preferred option. An easy way to predict the
weight of these engines based only on their power rating was necessary to be
found. Since no such method was available in the literature, an equation was
derived based on civil turboshaft/turboprop specifications available online
(Meier, 2005). A significant number of engines used in airborne applications
have been included in this study to derive the required equation. Figure 90
presents the weight and shaft power of the machines included in this study,
whilst the manufacturer and the model being used at are summarised in
Appendix A.3.
175
Figure 90 Weight vs. Shaft Power of Turboshaft/turboprop engines
Equation (6-1) was used in the Simulink models later being described. It should
be noted that only engines of rating less than 4MW were included in this study
both for reasons of specifications’ availability and also because this will be the
upper limit of rating for the DEAP engines.
𝐸𝑛𝑔𝑖𝑛𝑒′𝑠 𝑤𝑒𝑖𝑔ℎ𝑡 = (0.2338 ∗ 𝑆ℎ𝑎𝑓𝑡 𝑃𝑜𝑤𝑒𝑟) + 20.934 (6-1)
In all the proposed configurations of HEDP, superconducting machines were
assumed. Their weight was calculated based on the models being described in
4.2 and were also used during Chapter 5 calculations. Apart from the output
power, the RPM was also used as an input for these models. Values for the
speed of these machines were based on the DEAP project cases where speeds
around 12300 and 11100 RPM were assumed for the superconducting motors
and generators respectively (Wright et al., 2015). The weight of the cryo-coolers
required for these superconducting machines was estimated based on the
models described in the previous chapter and presented in Appendix A.2. This
method could give some misleading results since these machines most likely
will not have dedicated cryo-coolers but this weight estimation is used more as
y = 0.2338x + 20.934 R² = 0.9556
0
100
200
300
400
500
600
700
800
900
0 500 1000 1500 2000 2500 3000 3500 4000
We
igh
t (k
g)
Shaft Power (kW)
176
an indication on the relative weight of the cryo-cooler between the several
cases rather than an absolute number in each case.
Finally, the weight of the energy storage mechanisms will also be calculated
using Matlab/Simulink models. These models will take into account both the
energy and the power density of the candidate energy storage solutions. Table
23 summarises the battery technologies investigated in this study, whilst a
single 2035 prediction for the supercapacitors case was used. SMES was not
included in the models because of the uncertainty of the actual power and
energy density that they will present in the overall system. Using the values
available in the literature was considered as an “unfair” representation of their
case which might become attractive mainly due to the superconducting nature
of the whole network.
6.2.3 HEDP proposed configurations
A number of different architectures will be proposed and a weight sensitivity
study will be carried out for each case. The ideal energy storage system in
terms of weight will also be decided in each configuration.
Case 1: Use of energy storage during take-off
The first case is based on the E-thrust concept investigated by EADS and Rolls
Royce in the recent years. This concept was considered as a hybrid electric
propulsion system aiming to reduce fuel consumption and emissions of the next
generation aircraft. E-thrust implemented the Distributed Propulsion approach
consisted of six electrically driven fans powered either by an advanced battery
system or by a gas power unit depending on the phase of the flight (Singh,
2013). In the present study a superconducting version of this concept will be
explored. The energy storage will be used mainly during take-off reducing the
power requirements of the gas turbine. The latter will be designed to perform in
a constant power rating which will correspond to the power requirements during
cruise. The advanced battery system will assumed to be fully charged in the
beginning of each flight and it will be used as a boost for the take-off phase. It
will later been recharged during cruise (power derived from the GT) so that it
177
will be fully charged again for the next flight. The Simulink model developed to
estimate the weight of the energy storage system can be seen in Figure 91.
Figure 91 HEDP case 1 energy storage sizing Simulink model
The mission power demand as seen in Figure 89 is being used as the main
input. A code developed in MATLAB -being presented in Appendix A.4- gives
the required power at each moment of the fight. This mission profile is based on
the requirements of the baseline aircraft investigated during the DEAP project.
The maximum power of the turbo-generator (i.e. GTA) is also given as an input
combined with maximum and minimum power that the energy system could
produce. In order to size the latter, the battery energy capacity and its charge
rate limit were also used as inputs to the model. The overall target of the model
is to ensure that the mission power demand is being satisfied at all times by the
engine and the battery pack. A check of static upper bound block is being used
to ensure that there is never a power deficit in the system. The energy capacity
and power of the energy storage system is manually varied aiming on keeping
the minimum possible required weight for the whole network whilst the mission
demand is always satisfied. Figure 92 demonstrates the power output of the
GTA (red line) and the Energy Storage Device (blue line) during the whole flight
mission. The sum of these two lines corresponds to the DEAP aircraft mission
power demand earlier presented (Figure 89).
178
Figure 92 Case 1 GTA (red line) and Energy Storage (blue line) power output with
time
The main concept developed in this first case is to use the energy storage
during take-off in order to reduce the power requirements of the GTA.
Moreover, the GT can be optimised to run at an almost constant power
throughout the whole mission. The engine is slightly oversized in order to deal
with the additional power demand of the required cryo-coolers. The State of
Charge (SoC) of the energy storage mechanism in the first case examined in
this chapter is shown in Figure 93. As it was described earlier the battery
sharply discharges during take-off and it is recharged during cruise so that it will
be fully charged and ready to be used again for the next flight.
179
Figure 93 Case 1 Energy Storage State of Charge (SoC) in kJ with time (s)
Especially for this first case the sizing values for the engine and the battery
required for the baseline aircraft can be found in the following table:
Table 25 Case 1 GTA and Energy Storage sizing factors
Variable Value Units
GTA Power Output (1 engine) 3700 kW
Energy Storage Power Output 5000 kW
Energy Storage Maximum Energy 2.73 GJ
Energy Storage Charging Power 185 kW
The weight of the several energy storage options was calculated based on their
energy and power density capabilities via the Simulink subsystem demonstrated
in Figure 94. The four different battery technologies described in 6.1.1 and the
supercapacitor option were the technologies investigated.
180
Figure 94 Energy storage weight estimation Simulink model
The following table summarises the results of the Simulink model. Li-sulphur
technology seems to be the most attractive option in terms of weight for the
Case 1 configuration. For every technology apart from the Li-air battery type,
specific energy is the sizing restricted factor. It also becomes clear that the use
of a supercapacitor in such a configuration will be prohibitive.
181
Table 26 Case 1 Energy storage technology sizing values
Technology Sizing Restriction Mass (kg)
Li-ion (SOTA) Energy 3795
Li-ion (future) Energy 1897
Li-air Power 7143
Li-sulphur Energy 1518
Supercapacitor (future) Energy 15180
Energy storage was not the only component that adds significant weight in this
first proposed conceptual aircraft architecture. Using equation (6-1) the
anticipated weight of the turboshaft engine was estimated. The superconducting
machine models were used to estimate the weight of the generator driven by
the main engine, whilst the required cryo-cooler weight was also calculated by
assuming 99.9% efficiency for the machines and an operational temperature of
20K. The weight of the superconducting motors was not included in this study
because it is not expected to differ between the various configurations. The
required propulsive power will be relatively constant in all three cases and thus
the weight of the propulsors (superconducting motor driven fans) is expected to
remain the same. Figure 95 presents the overview of the Simulink model used
for the first case, whilst Table 27 summarises the weight of each main
component of this configuration. All the subsystems have been described earlier
in the research study and will not be analysed any further.
182
Figure 95 Case 1 Electric Components sizing Simulink model
The following table will be used as a comparison with the rest of the cases
being studied in this section.
Table 27 Summary of Case 1 components’ weight
Component Unit Mass (kg) Qty. Total
Mass (kg)
Turboshaft Engines 886 1 886
Generators 117.4 1 117.4
Cryo-cooler 194.2 1 194.2
Energy Storage 1518 1 1518
Total System’s Weight 2715.6
183
Case 2: Use of Energy storage during cruise
Battery powered aircraft are growing in interest mainly due to the important fuel
and emission benefits that could potentially offer. However, their specific energy
is significantly lower than the one of kerosene and depending on the range of
the aircraft could lead to enormously heavy battery packs. Besides, the case of
supercapacitors is not even considered in such an application due to their low
specific energy. There have been many studies investigating concepts similar to
the one of Case 2. The Boeing’s Sugar Volt aircraft (1.2.2) for example is using
batteries to enable portions of flight with low emissions. More specifically, twin
engines are jet fuel powered during take-off, whilst at altitude the hybrid/electric
system takes over (Owano, 2012). The all electric transport concept was also
investigated by EADS (later renamed to AGI) with the Voltair conceptual design
investigating the feasibility of such an aircraft (Stuckl, Van Toor and
Lobentanzer, 2012). Finally, Bauhaus Luftfahrt also examined the so-called
Universally-Electric Systems architecture where advanced Li-ion batteries
constituted the only electric power source (Isikveren et al., 2012).
In this second case, the advanced battery system is being used exclusively
during cruise. Since the cruise phase is the lengthiest one, significant fuel burn
and emissions benefits are expected by using only battery-powered propulsion.
These benefits will not be quantified during this study, where only a weight
feasibility study will be carried out for an aircraft of the range similar to the
DEAP baseline example. The Simulink model being used for this case is shown
in Figure 96, where the main difference is the use of a second twin engine
during take-off and landing phases instead of the energy storage devices which
in this case take over during the cruise at altitude.
184
Figure 96 HEDP case 2 energy storage sizing Simulink model
A similar sizing strategy to the one presented in Case 1 is also used for Case 2.
However, this time the energy storage is being sized so that it can produce the
required power during the cruise phase, whilst the twin engines are used for the
demanding take-off, climb, and landing phases. Table 28 summarises these
sizing factors for the engines and the energy storage devices. It is clear that the
required maximum energy of the energy storage system will have an enormous
effect on the size of these devices.
Table 28 Case 2 GTA and Energy Storage sizing factors
Variable Value Units
GTAs Power Output (2 engines) 4300 kW
Energy Storage Power Output 3350 kW
Energy Storage Maximum Energy 52.4 GJ
Energy Storage Charging Power 185 kW
The power output values mentioned in the table above can also be seen in
Figure 97. The red line represents the output power of the two GTAs who are
being used in full power during the take-off and almost half power during the
descent, landing and taxi phases.
185
Figure 97 Case 2 GTAs (red line) and Energy Storage (blue line) power output
(kW) with time (s)
In case 2 the energy storage system is assumed to be fully charged in the
beginning of the flight, whilst it is fully discharged when the aircraft will be
landing. This tactic creates several issues; firstly, the batteries will have to be
recharged again before the next flight. This could take several minutes or hours
making this choice as impractical. Depending on the cost another option will be
the complete replacement of the batteries with new fully charged ones. Another
issue will be the Depth of Discharge (DoD) of these batteries. Typically,
batteries’ life expectancy highly depends on their DoD levels. 100% DoD
(battery completely empty) is really harmful for their life expectancy. Lower
values of DoD could increase their lifetime but also increase the weight of these
devices. The following figure shows the SoC of the energy storage system in
Case 2:
186
Figure 98 Case 2 Energy Storage State of Charge (SoC) in kJ with time (s)
Table 29 indicates the weight of several energy storage technologies in a
configuration similar to the universally all electric aircraft investigated in Case 2.
Li-air technology seems the preferable choice in terms of weight this time.
Nonetheless, the energy storage system will still weigh almost 20 tonnes even
in the best case scenario. This added weight is unacceptable for an aerospace
application.
Table 29 Case 2 Energy storage technology sizing values
Technology Sizing Restriction Mass (kg)
Li-ion (SOTA) Energy 72840
Li-ion (future) Energy 36420
Li-air Energy 19420
Li-sulphur Energy 29130
Supercapacitor (future) Energy 291300
187
Such a configuration does not seem feasible at least with the assumptions
being made in this research study. Table 30 proves the infeasibility of this
concept where a total weight of around 22 tonnes is estimated for such a
configuration. Case 2 does not bring any weight benefits for the weight of the
electrical components involved, however, it offers some benefits that are difficult
to be quantified at this stage. Apart from the obvious fuel burn benefits during
cruise, the optimisation of the engines for the take-off phase would increase the
efficiency of these GTs and consequently reduce the fuel burn even further.
Noise and emissions reductions will also be significant gains of such
architecture.
Table 30 Summary of Case 2 components’ weight
Component Unit Mass (kg) Qty. Total
Mass (kg)
Turboshaft Engines 1026 2 2052
Generators 134.9 2 269.8
Cryo-cooler 294.8 1 294.8
Energy Storage 19420 1 19420
Total System’s Weight 22036.6
In this study a maximum specific energy value of 750 Wh/kg has been assumed
for the battery technologies under investigation. However, in similar studies
more optimistic energy density values up to 2000 Wh/kg have been assumed by
several companies and institutions (Isikveren et al., 2012). Figure 99
demonstrates the effect that specific energy assumptions have to the weight of
the battery packs in a configuration similar to the one described in Case 2. The
optimistic assumption of 2000 Wh/kg could save up to 12.095 tonnes to the
overall weight of the architecture under consideration.
188
Figure 99 Weight of the battery system vs. specific energy assumptions for Case
2 configuration
Generally, it can be seen that the relationship between specific energy and
overall weight of the battery system is not linear as one might think. A more
exponential trendline can be noticed, where improvements of specific energy
values higher than 1300 Wh/kg do not affect the overall weight of the system to
the same extend as they do in the range of 700-1300 Wh/kg.
It is becoming clear that fully electric (battery-powered) configurations for
aircraft of similar range to the DEAP one are too heavy to be realised. Possibly,
in aircraft of smaller range and minimum power requirements this concept could
become feasible. Especially, if the optimistic targets of future advanced battery
systems of energy density higher than 1300 Wh/kg could be met.
Case 3: Use of Energy storage as a supplementary power unit source
In this third choice energy storage would play a more secondary but still
important role. The idea behind this case is again to design engines of a
“specific purpose”. The GTs will be optimised for the cruise phase, whilst the
deficit in required power demand during the take-off and climb phases will be
covered by the energy storage subsystem. This case is a modification of
SUGAR Volt conceptual design developed by Boeing and presented in the
0
5
10
15
20
25
0 500 1000 1500 2000 2500
Wei
ght
(to
nn
es)
Specific Energy (Wh/kg)
189
Introduction Chapter (1.2.2). In this conceptual configuration one of the two
engines could be switched off during cruise increasing the fuel, emissions, and
maintenance cost benefits. The switched-off engine could also be altered from
flight to flight increasing their life expectancy and reducing their maintenance
costs.
Table 31 summarises the required power ratings for the GTAs and the energy
storage mechanisms in Case 3. The twin engines are specifically sized so that
one of them would be enough to deal with the cruise power requirements. The
energy storage is sized to provide the additional required power during take-off
and climb or any other dynamic requirement of the whole mission (i.e. landing,
emergency cases).
Table 31 Case 3 GTA and Energy Storage sizing factors
Variable Value Units
GTAs Power Output (2 engines) 3820 kW
Energy Storage Power Output 1000 kW
Energy Storage Maximum Energy 0.464 GJ
Energy Storage Charging Power 185 kW
The Simulink model of this case is similar to the one presented in Case 2. The
main difference is the power outputs of the several components throughout the
whole mission. These can be seen in Figure 100 where one of GTAs is being
used for the whole mission, whilst the second GTA is switched off during cruise,
descent, and landing flight phases. On the other hand, energy storage is sized
so that it can provide the extra required power during take-off and landing
phases. Both the engines and the energy storage devices are slightly oversized
to deal with the requirements of the cooling system.
190
Figure 100 Case 3 GTA 1 (red line), GTA 2 (green line) and Energy Storage (blue
line) power output (kW) with time (s)
The energy storage is assumed to be fully charged in the beginning of the flight.
During cruise the GTA is providing enough energy to recharge the batteries,
while the landing power requirements are lasting for a short period of time that
do not practically affect the SoC of the energy storage subsystem. By the end of
the whole mission, batteries are again fully charged and ready for the next flight.
The following figure demonstrates the SoC of the energy storage system in
Joules for the Case 3 of this section.
Figure 101 Case 3 Energy Storage State of Charge (SoC) in kJ with time (s)
The preferable energy storage technology in terms of weight is again estimated
via the Simulink submodel presented in Figure 94. In the Case 3 study, Li-
sulphur battery technology is proved to be the most lightweight choice adding to
191
the system only 258 kg. On the other hand, Li-air technology that seemed to be
the ideal option for Case 2 weighs significantly more (~1430kg) mainly because
of its anticipated specific power. This technology is the only one that specific
power is the decisive sizing factor instead of specific energy which dictates the
weight of the rest of energy storage technologies. This proves that an exclusive
look on the energy density values of future energy storage trends could lead to
misleading results. Table 32 sums up the mass of the energy storage
technologies under investigation for Case 3.
Table 32 Case 3 Energy storage technology sizing values
Technology Sizing Restriction Mass (kg)
Li-ion (SOTA) Energy 645
Li-ion (future) Energy 322.5
Li-air Power 1429
Li-sulphur Energy 258
Supercapacitor (future) Energy 2580
In Case 3 the estimated weight of energy storage does not seem to affect the
feasibility of the conceptual design. On the contrary, it is one of the most
lightweight components of the whole electric system. Table 33 summarises the
weight of all the components involved in this study. The overall system’s weight
is less than 3 tonnes and even lighter than Case 1 configuration. Besides, this
configuration has a comparable weight to a more conventional architecture (no
energy storage involved) because the added weight of energy storage system is
compensated by the expected weight reduction of the GTAs and the cooling
system. More specifically, a configuration for the DEAP aircraft with two GTAs
and no energy storage involved will weight 2623.5kg, almost 100kg less than
Case 1, but 20 kg heavier than Case 3 configuration.
192
Table 33 Summary of Case 3 components’ weight
Component Unit Mass (kg) Qty. Total
Mass (kg)
Turboshaft Engines 914 2 1828
Generators 120.9 2 241.8
Cryo-cooler 277.9 1 277.9
Energy Storage 258 1 258
Total System’s Weight 2605.7
Case 3 combines the concepts of distributed propulsion, superconductivity, and
novel flight cycles in the most efficient way in terms of weight. Furthermore, the
fact that GTAs are sized for a specific phase of the mission (i.e. cruise) frees
the design approach of the engines from take-off power requirements, possibly
making their optimisation a simpler procedure. Hence, efficiency benefits are
expected to add up to the already estimated weight profits.
6.2.4 Final Remarks
In the previous section three different cases of HEDP configurations with energy
storage were explored in terms of weight. The first case investigated the use of
energy storage as the main power source during take-off, whilst the whole
network also included a single GTA as the main prime mover during the rest of
the flight mission. Case 2 was based on the idea of a battery-powered aircraft
where battery packs are used as the main propulsion power source during
cruise. Finally, a more conservative use of energy storage (as a “boost power
unit”) was investigated in Case 3 where energy storage systems are mainly
used to secure the constant function of the GTAs during the whole mission.
Case 1 and 3 proved to be competitive configurations compared to similar
HEDP architecture where no use of energy storage will be made. Figure 102
demonstrates a weight comparison between the various components of the
network for Cases 1, 3, and a hybrid DP configuration without any energy
storage in use. Case 2 was not included in this comparison since the weight of
193
energy storage and consequently of the whole network is around 20 tonnes
heavier, thus no useful comparison could be made at this stage.
Figure 102 Overall Weight comparison for the different cases
6.3 Sensitivity Study for Hybrid Configurations for aircraft of
different sizes
The extension on different aircraft sizes was problematic due to the uncertainty
of the mission profile of different aircraft implementing the HEDP concept. It is
really challenging to accurately compare and contrast the thrust rating of a jet
engine with the power rating of the turboshaft engines potentially used in a
HEDP configuration. Such a comparison could lead to misleading results as
these two quantities are not equivalent. In the distributed propulsion concept the
engines will be rated by how much power they will need to deliver to the
propellers (i.e. motor-driven fans). In a conventional “jet engine configuration”,
the propulsive power of the engine is decided using the following equation:
𝑃 = 𝐹𝑑
𝑡
(6-2)
Power (P) is the force (F) needed to drive an item over a distance (d) divided by
the time (t). In a jet aircraft this force is equal to the thrust produced by the
engines, hence:
0
500
1000
1500
2000
2500
3000
Engines Generators Cooling System Energy Storage Total Weight
Wei
ght
(kg)
Case 1 Case 3 No Energy Storage
194
𝑃 = 𝑇 ∗ 𝑣 (6-3)
However, in the HEDP cases turboshaft engines will be used instead of
turbojets. In order to decide the required mission power of each reference
aircraft a similar strategy to the one being used in Chapter 5 will be followed. A
value of power demand per passenger will determine the required power of the
most important flight phases (i.e. SLS, EoR, Climb, and Cruise). Landing,
descent, and taxi phases are typically being determined as a percentage of the
cruise power demand (140, 80, and 40 % respectively). In regards to the time
length of the flight mission of each aircraft this will be estimated based on the
maximum range and speed of each reference aircraft. Only Cases 1 and 3 will
be investigated in this section. Case 2 has already been proven significantly
heavy for an aircraft of the size of DEAP baseline (short to medium range). As
the range and power requirements increase, the required weight for the battery
packs will be expected to increase even further enhancing the infeasibility of the
whole “Case 2” concept.
6.3.1 Reference Aircraft Mission Profiles
The previous subchapter has shown some of the potential benefits that a hybrid
configuration could offer in a short to medium range aircraft. It seems
reasonable to extend this study to a vast variety of aircraft with different power
requirements and ranges. The same reference aircraft used in Chapter 5.4.1
will also be the reference aircraft models for this chapter, with the exception of
Bombardier CRJ-1000 which was substituted with a smaller model of the same
company (i.e. CRJ-100) and without the B777 example. This choice of aircraft
was based on the idea of exploring a wide range of mission profiles in terms of
power required and length. Furthermore, the consistency of the whole study and
the investigation of the state of the art and most popular aircraft models were
other reasons for this choice. Table 34 summarises the main characteristics of
these reference models:
195
Table 34 Reference Aircraft Main Characteristics
Model PAX Range (km) Speed
(km/h)
Cruise
Power (kW)
Cruise
time (min)
CRJ-100 50 1800 860 1673 126
DEAP 100 3704 918 3346 242
B737 189 5080 1000 6324 304
B787 242 14500 950 8098 915
A350 475 14800 945 15894 940
A380 700 15000 945 23422 952
The PAX, Range, and Speed columns were filled based on data found in the
literature (www.airlines-inform.com, 2012). The cruise time was calculated using
the following equation:
𝐶𝑟𝑢𝑖𝑠𝑒 𝑡𝑖𝑚𝑒 (min) = (𝑅𝑎𝑛𝑔𝑒
𝑆𝑝𝑒𝑒𝑑) ∗ 60
(6-4)
The duration of the rest of the flight cycles was assumed to be the same as the
DEAP baseline aircraft. Slight changes on the duration of each flight phase
might occur but for reasons of simplicity these have been ignored. Table 34 only
includes the cruise power requirements of each aircraft. However, for every
phase of the reference aircraft the following set of equations has been used:
𝑆𝐿𝑆 𝑃𝑜𝑤𝑒𝑟 (𝑘𝑊) = 86.18 ∗ 𝑃𝐴𝑋 (6-5)
𝐸𝑂𝑅 𝑃𝑜𝑤𝑒𝑟 (𝑘𝑊) = 85.78 ∗ 𝑃𝐴𝑋 (6-6)
196
𝐶𝑙𝑖𝑚𝑏 𝑃𝑜𝑤𝑒𝑟 (𝑘𝑊) = 38.16 ∗ 𝑃𝐴𝑋 (6-7)
𝐶𝑟𝑢𝑖𝑠𝑒 𝑃𝑜𝑤𝑒𝑟 (𝑘𝑊) = 33.46 ∗ 𝑃𝐴𝑋 (6-8)
The constants in the equations above have been estimated using the DEAP
aircraft power requirements per phase and per passenger. This might seem as
a simplistic method to calculate the power requirements of each aircraft but
even in the conventional configurations the thrust requirements of the main
engines change linearly with the number of passengers as can be seen in
Figure 103. There is no reason to believe that this would be any different for a
HEDP configuration.
Figure 103 Main engines’ thrust ratings vs number of passengers in the
reference aircraft
For the remaining parts of the flight the required power was estimated based on
the cruise power requirements and by using the following equations:
𝐷𝑒𝑠𝑐𝑒𝑛𝑡 𝑃𝑜𝑤𝑒𝑟 (𝑘𝑊) = 0.8 ∗ 𝐶𝑟𝑢𝑖𝑠𝑒 𝑃𝑜𝑤𝑒𝑟 (6-9)
0
50000
100000
150000
200000
250000
300000
350000
0 100 200 300 400 500 600 700 800
Thru
st R
equ
irem
ents
(lb
)
PAX
197
𝐿𝑎𝑛𝑑𝑖𝑛𝑔 𝑃𝑜𝑤𝑒𝑟 (𝑘𝑊) = 1.4 ∗ 𝐶𝑟𝑢𝑖𝑠𝑒 𝑃𝑜𝑤𝑒𝑟 (6-10)
𝑇𝑎𝑥𝑖𝑖𝑛𝑔 𝑃𝑜𝑤𝑒𝑟 (𝑘𝑊) = 0.4 ∗ 𝐶𝑟𝑢𝑖𝑠𝑒 𝑃𝑜𝑤𝑒𝑟 (6-11)
Using all the aforementioned information the mission profile of each aircraft was
estimated. Figure 104 combines the mission profiles of all five reference aircraft
in one graph. From the really short range example of CRJ-100 to the biggest
aircraft currently in service (A380) a weight assessment of hybrid configurations
will be presented in the following section.
Figure 104 Mission Profiles of the reference aircraft
6.3.2 Results and comments
As it was mentioned before only Cases 1 and 3 of 6.2.3 will be included in this
sensitivity study since Case 2 (i.e. use of energy storage during cruise) has
already been proven infeasible even for a short range aircraft such as the DEAP
aircraft. A weight comparison between configurations with and without energy
storage will be carried out for each aircraft separately, whilst an overall
comparison will conclude this section.
198
Bombardier CRJ-1OO
The present aircraft was selected as an example of a really short range aircraft
that could benefit by the use of energy storage to implement novel flight cycles
which could increase the efficiency and reduce the costs (fuel, maintenance
etc.) of future hybrid architectures. Table 35 sums up the mass of different
energy storage technologies used in an aircraft similar to CRJ-100 for the
hybrid/electric configurations of Cases 1 and 3. Note that for both cases the
engines’ power output was assumed to be 1700 kW (single engine for Case 1
and two identical engines 1700 kW each for Case 3).
Table 35 CRJ-100’s Energy storage technology sizing values
Technology Sizing
Restriction
(Case 1)
Mass (kg)
(Case 1)
Sizing
Restriction
(Case 3)
Mass (kg)
(Case 3)
Li-ion (SOTA) Energy 2224 Energy 756.2
Li-ion (future) Energy 1112 Energy 378.1
Li-air Power 3857 Power 1314
Li-sulphur Energy 889.6 Energy 302.5
Supercapacitor
(future)
Energy 8896 Energy 3025
The table above shows that Li-sulphur seems to be the ideal option in terms of
weight. However, future improvements in Li-ion technology could create
competitive products that might be preferable due to the maturity and better
understanding of this technology. Notwithstanding other criteria, Li-sulphur
sizing values will be used in the following sizing study of CRJ-100 example.
Furthermore it should be stated that the power requirements for the energy
storage mechanisms was 2700 and 920 kW for Case 1 and 3 respectively.
Finally, the maximum energy capacity was 1.60 GJ for the Case 1 configuration
and 0.544 GJ for Case 3.
199
Table 36 Summary of CRJ-100 components’ weight
Component Mass (kg)
(case 1)
Mass (kg)
(case 3)
Mass (kg)
(No E.S.)
Turboshaft Engines 418.4 836.8 1049
Generators 57.17 114.34 142.34
Cryo-coolers 133 186.3 209.3
Energy Storage 889.6 302.5 -
Total System’s Weight 1498.17 1439.94 1400.64
The above table shows that the hybrid/electric configurations are actually weight
neutral in the case of CRJ-100 aircraft size. However, it should be noted that
the novel hybrid flight cycles of Case 1 and 3 could offer extra benefits that
cannot be easily quantified in these preliminary studies. Fuel consumption,
noise levels, emissions, and maintenance costs will all decrease in the case of
hybrid architectures. The option of hybrid/electric aircraft in case of a short
range aircraft such as CRJ-100 and DEAP aircraft seems to be an attractive
and viable option for the next aircraft generations at least based on the current
assumptions for future energy and power densities of energy storage
mechanisms.
Boeing 737
Although the feasibility of these novel flight cycles with energy storage use
seems a rather beneficial alternative for the short range aircraft, it is interesting
to investigate their viability for medium range aircraft. In this category Boeing’s
737 is a representative example. For B737 Case 1 a 6.6 MW GT has been
used, whilst two 6.8 MW engines were assumed for Case 3 example. This led
to energy storage power requirements of 9.7 and 2.75 MW respectively in order
the power demands to be satisfied at all flight phases of the mission in both
Cases. The following table presents the anticipated mass of the energy storage
mechanisms under investigation:
200
Table 37 B737‘s Energy storage technology sizing values
Technology Sizing
Restriction
(Case 1)
Mass (kg)
(Case 1)
Sizing
Restriction
(Case 3)
Mass (kg)
(Case 3)
Li-ion (SOTA) Energy 8201 Energy 1946
Li-ion (future) Energy 4101 Energy 973
Li-air Power 13860 Power 3929
Li-sulphur Energy 3280 Energy 778.4
Supercapacitor
(future)
Energy 32800 Energy 7784
It should be noted that maximum energy capacities of 5.9 and 1.4 GJ were
required for Case 1 and Case 3 battery options. Once again Li-sulphur was
found to be the lightest option. For Case 3 future Li-ion technology could also
be used without a severe weight penalty.
Table 38 summarises the components’ weight for the hybrid and more
conventional configurations of B737 type of aircraft. This time Case 1 weighs
almost half a tonne more than the other two versions. The use of energy
storage as a supplementary power unit during take-off and landing phases
(Case 3) gives a slight weight penalty of 40kg for the medium range example of
B737 type of aircraft. This however will be counteracted by the anticipated fuel
and emission benefits described earlier in this study.
201
Table 38 Summary of B737 components’ weight
Component Mass (kg)
(case 1)
Mass (kg)
(case 3)
Mass (kg)
(No E.S.)
Turboshaft Engines 1564 3222 3850
Generators 200.05 412.2 487
Cryo-coolers 258.3 371.2 406.8
Energy Storage 3280 778.4 -
Total System’s Weight 5302.35 4783.8 4743.8
It is becoming clear that if the distributed propulsion will indeed be chosen as
the way forward in the next generation aircraft, energy storage could play an
important role further enhancing the benefits of such a configuration at least for
the short to medium range aircraft models. The remaining reference aircraft will
investigate the viability of the concept for the longer range aircraft that are
currently dominating the air traffic (i.e. B787, A350, and A380).
Boeing 787
Boeing’s 787 aircraft has already been presented in detail during the previous
sections of this research study as being the most representative example of
MEA. The electric network of this aircraft consists of state of the art electrical
components. Further electrification of this type of aircraft is anticipated and the
proposed configurations would be an important step towards this direction.
In order the mission profile of such an aircraft to be satisfied at all times, a 8.4
MW GTA was chosen for the Case 1 configuration and two GTAs with power
rating of 8.5 MW each were used for the Case 3 B787 design. Table 39
presents the estimated mass for the energy storage mechanisms in an aircraft
similar to the “Dreamliner” example. Battery packs of 12.5 and 4 MW were
required for Case 1 and 3 respectively. The energy capacities of these two
energy storage subsystems were 7.7 and 2.04 GJ and were the sizing
202
restriction factor for all the cases but the Li-air battery which was sized based
on the power requirements.
Table 39 B787‘s Energy storage technology sizing values
Technology Sizing
Restriction
(Case 1)
Mass (kg)
(Case 1)
Sizing
Restriction
(Case 3)
Mass (kg)
(Case 3)
Li-ion (SOTA) Energy 10700 Energy 2836
Li-ion (future) Energy 5352 Energy 1418
Li-air Power 17860 Power 5714
Li-sulphur Energy 4281 Energy 1134
Supercapacitor
(future)
Energy 42810 Energy 11340
Li-sulphur technology is again the most lightweight option and especially for the
Case 1 configuration seems to be the ideal technology in terms of weight. It is
important to note that if energy density was the only factor under investigation,
then Li-air battery would have been almost two times lighter than the Li-sulphur
equivalent. Hence, if these batteries could be designed in a more efficient way
in terms of power density (probably at a slight expense of energy density limits)
a more competitive Li-air product could be designed.
The list of the components’ weight in the 787 type of aircraft can be seen in
Table 40. The difference in overall weight between Case 1 concept and the
other two designs starts to grow. Although Case 1 involves fewer components
the scalability of the energy storage subsystem creates a heavier overall
network. On the other hand, Case 3‘s configuration remains competitive in
terms of weight with a distributed propulsion design with no energy storage in
place.
203
Table 40 Summary of B787 components’ weight
Component Mass (kg)
(case 1)
Mass (kg)
(case 3)
Mass (kg)
(No E.S.)
Turboshaft Engines 1985 4016 4918
Generators 250.6 506.6 612.2
Cryo-coolers 291.3 415.8 461.7
Energy Storage 4281 1134 -
Total System’s Weight 6807.9 6072.4 5991.9
Airbus A350
The last two reference aircraft are the most demanding in terms of power and
range. The latest model being launched by Airbus is the A350 aircraft which
according to equations (6-5) and (6-8) will require around 40 and 16 MW of
power during the take-off and cruise phase respectively. Thus, the engines
power rating was accordingly set to 16 and 16.2 MW for the two cases leading
to energy storage power demand of 25 (Case 1) and 8.6 (Case 3) MW. Table
41 highlights the derived weight of the various energy storage alternatives.
204
Table 41 A350‘s Energy storage technology sizing values
Technology Sizing
Restriction
(Case 1)
Mass (kg)
(Case 1)
Sizing
Restriction
(Case 3)
Mass (kg)
(Case 3)
Li-ion (SOTA) Energy 22240 Energy 6297
Li-ion (future) Energy 11120 Energy 3148
Li-air Power 35710 Power 12290
Li-sulphur Energy 8896 Energy 2519
Supercapacitor
(future)
Energy 88960 Energy 25190
It is no surprise that Li-sulphur is again the lightest energy storage option for
both configurations. However, the almost 9 tonnes added in Case 1 study might
prove to be a prohibitive number. It is also interesting to point out that if current
technology was to be used for these novel designs then more than 22 tonnes of
Li-ion battery would be needed in Case 1. This shows how essential the
technology improvements are for the feasibility of these concepts.
The following table compares the weight of the components for aircraft sizes
similar to A350 for the two novel configurations using energy storage (Cases 1
and 3) as well as for a superconducting distributed propulsion version (no
energy storage) of such an aircraft.
205
Table 42 Summary of A350 components’ weight
Component Mass (kg)
(case 1)
Mass (kg)
(case 3)
Mass (kg)
(No E.S.)
Turboshaft Engines 3762 7616 9612
Generators 454.8 920 1142.2
Cryo-coolers 403.1 580.2 656.1
Energy Storage 8896 2519 -
Total System’s Weight 13515.9 11635.2 11410.3
In this case the version with no energy storage is again the lighter configuration,
with more than 200 kg difference to the third case design option. On the other
hand, the use of energy storage during take-off in a one engine configuration
(Case 1) is almost two tonnes heavier, a number that cannot be neglected. As
the range and size of the aircraft increases, the weight difference between Case
1 and the other two options rises accordingly.
Airbus A380
The last reference aircraft is also the largest aircraft currently in service. Its
power demand reaches the 60 MW range during take-off and the 23.5 MW
during cruise. Following the same strategy as in the previous examples the
energy storage required power was found to be 36.6 MW for the Case 1 model
and 12.4 MW for the Case 3 design option. Hence, the required power output in
Case 1 is almost three times higher than in Case 1, a trend that was observed
in almost every reference aircraft being investigated in this section. An energy
storage mass summary of the different technologies for both cases can be
found in the following table.
206
Table 43 A380‘s Energy storage technology sizing values
Technology Sizing
Restriction
(Case 1)
Mass (kg)
(Case 1)
Sizing
Restriction
(Case 3)
Mass (kg)
(Case 3)
Li-ion (SOTA) Energy 32250 Energy 9091
Li-ion (future) Energy 16120 Energy 4545
Li-air Power 52290 Power 17710
Li-sulphur Energy 12900 Energy 3636
Supercapacitor
(future)
Energy 129000 Energy 36360
Almost 13 tonnes of energy storage mechanism will be required in a Case 1
configuration even in the best case scenario. This is a value that at first glance
seems prohibitive for an aerospace application. Li-sulphur battery, as in the rest
examples, is the most promising technology for a long range aircraft such as
A380. Just less than one tonne heavier is the Li-ion technology based on the
optimistic future predictions for the Case 3 conceptual design.
Table 44 Summary of A380 components’ weight
Component Mass (kg)
(case 1)
Mass (kg)
(case 3)
Mass (kg)
(No E.S.)
Turboshaft Engines 5585 11264 14148
Generators 656.6 1323.4 1635
Cryo-coolers 494.2 713.9 806.9
Energy Storage 12900 3636 -
Total System’s Weight 19635.8 16937.3 16589.9
The table above sums up the components’ weight for the last reference aircraft.
Case 1 is this time more than three tonnes heavier than a configuration with no
207
energy storage in use. The latter is also 400 kg lighter than the Case 3 design
option which however can still be considered as a competitive alternative option
in terms of weight.
6.3.3 Final Remarks
In this section the feasibility in terms of weight of novel hybrid configurations
using energy storage was investigated. The study had two main targets: firstly
to decide which energy storage technology is the most promising in terms of
weight for hybrid configurations such as the ones described in 6.2 and secondly
to explore if the use of energy storage either as the main power unit during
take-off (Case 1) or as a supplementary power unit in a two engines
configuration (Case 3) are feasible design concepts for a wide range of aircraft.
In regards to the first target, Li-sulphur was the clear winner compared to the
rest of energy storage options. This technology was estimated to be the most
lightweight option in every single aircraft (short or long range) for both cases.
Only future Li-ion batteries could potentially be competitive products for short to
medium range aircraft especially if their technology maturity is taken into
consideration. Furthermore, the cyclability of Li-ion batteries is significantly
higher than the Li-sulphur case, a fact that makes their option a lot more
attractive at least in regards to the cost. Unless future Li-S life cycles do not
improve significantly then Li-ion technology might be the preferable choice
especially in Case 3 configurations where the weight difference was not that
profound. Li-air technology, which by many has been considered as one of the
most promising battery technologies of the future, suffers from a low specific
power that restricts its size in most of the cases instead of the specific energy
factor. However, if higher values of specific power than the ones assumed in
this study could be obtained then this technology could become the ideal
candidate for these hybrid configurations. Nonetheless, Li-air technology life
expectancy is also a barrier in using this type of batteries in an aerospace
application.
An overall weight comparison of the system for each reference aircraft
implementing the three different novel configurations is demonstrated in Figure
208
105. For short to medium range aircraft (i.e. CRJ-100, DEAP, B737) all three
design options present comparable systems in terms of weight and further
sensitivity studies concentrating on different factors need to be carried out. On
the other hand, as the take-off power demand increases (i.e. B787, A350, and
A380) Case 1 starts to weigh significantly more than the other two versions. Any
other benefits that such a configuration could offer will most probably be met by
the Case 3 design option.
Figure 105 Weight Comparison between Case 1, Case 3, and a configuration
without energy storage for all the reference aircraft
It is also important to observe how the weight of the energy storage differs with
the power demand of each aircraft. Figure 106 demonstrates how the weight of
Li-sulphur batteries increases with the SLS power demand as we go from
smaller to larger aircraft (i.e. CRJ-100 to Airbus A380). In both cases an almost
linear relationship can be observed. However, in Case 1 the angle of this linear
trendline is a lot sharper revealing the main reason why Case 1 is less
competitive in terms of weight in the longer range aircraft. Similar trends could
be seen in every other energy storage technology, but Li-sulphur was chosen
as the most promising one of all.
0
5000
10000
15000
20000
25000
CRJ-100 DEAP B737 B787 A350 A380
Tota
l Wei
ght
(kg)
Case 1 Case 3 No Energy Storage
209
Figure 106 Li-sulphur weight vs. reference aircraft SLS power requirements
To sum up, this study showed that a hybrid configuration implementing the
superconducting distributed propulsion concept with extensive use of energy
storage could be feasible in the medium term future (i.e. 2035 timeframe). Li-
sulphur has proven to be the most promising technology in terms of weight for
almost all the proposed configurations and aircraft. Supercapacitors on the
other hand seem to be too heavy to be considered as a main power unit in
these designs. Furthermore, the concept of an all-electric aircraft which will be
using batteries as the main power source during cruise, at this point and with
the relatively conservative assumptions been made for the 2035 timeframe, has
been found significantly heavier than the rest of the conceptual designs. Finally,
although in the medium to long rage aircraft the use of energy storage as a
main power source during take-off is clearly not the preferred option in terms of
weight, this is not the case in shorter range aircraft similar to the DEAP baseline
aircraft. Hence, other factors such as fuel savings, emissions and noise
reductions, maintenance costs etc. need to be explored in more depth so that
the ideal architecture for each aircraft to be decided. This should be combined
with a more detailed sensitivity study for the energy storage options which will
include factors such as DoD, life cycles, and safety.
0
2000
4000
6000
8000
10000
12000
14000
0 10000 20000 30000 40000 50000 60000 70000
Ener
gy s
tora
ge w
eigh
t (k
g)
SLS Power Demand (kW)
Case 1
Case 3
210
This chapter proved the feasibility of hybrid configurations in terms of weight for
the majority of the cases. These proposed propulsion systems are attractive for
several additional reasons. First of all, the safety of these designs is enhanced
mainly due to the various power sources which will be available in the aircraft.
Both the GTs and the energy storage should be sized in such a way that they
will be able to deal with any possible safety case. The increase in critical
components also improves the reliability and redundancy of the whole system.
Finally, the flexibility of the whole network is enhanced. The concept behind
these proposed designs is the optimisation of each prime mover for a specific
function. This will have a direct positive effect on the efficiency of each
component and consequently in the emissions (i.e. noise, NOx, fuel
consumption) of the whole aircraft.
6.4 Key study Limitations
It is important to emphasise that this study could be used only as a preliminary
high TRL investigation of the feasibility in terms of weight for some novel hybrid
configurations. Thus, there are important limitations due to some uncertainty of
the assumptions been made but mainly because of the fact that only some
aspects of these configurations were investigated. More specifically:
The energy storage options review was exclusively focused on the
weight of these components. Other factors that could prove to be
equally important in the selection of the ideal energy storage mechanism
were not part of this study. Particularly, the cyclability and safety issues
of each type of battery could determine the final decision being made.
Volume is also an important attribute especially in the case of an
aerospace application. In large aircraft where the required power is
extremely high, the volume of the batteries could be the main barrier of
the whole concept. Also, an aerospace propulsion system is far more
sensitive in reliability issues than any other ground based application.
The recent examples of Boeing 787 battery issues confirm the need of a
highly reliable and safe energy storage mechanism particularly in the
cases when the latter is being used as the main power source during
211
take-off. Nonetheless, weight is still considered as the decisive factor in
the feasibility of the whole concept and that is the reason this study is
dedicated exclusively on it.
This study has been based on the fact that future aircraft propulsion
systems will adopt the distributed propulsion concept using a fully
superconducting network and cryo-coolers as their cooling system. A
weight comparison with a more conventional configuration similar
to the one being used from the majority of the reference aircraft could
have given some useful results. However, the whole purpose of this
chapter was to further explore the possibilities of TeDP configurations
and decide on the possible role that energy storage mechanisms could
play on these architectures.
The additional benefits of using energy storage in these hybrid
configurations were not explored. These include fuel consumption,
fuel weight benefits, noise levels, emissions reduction, and maintenance
costs. Although important, all these advantages are difficult to be
quantified in such an early stage. Nonetheless, it is reasonable to
believe that if the weight factor does not block the feasibility of these
conceptual designs then most probably these added benefits will
enhance the attractiveness of the whole concept.
The efficiency deficit of the GTs in the altitude and the EOR safety
case were not clearly taken into consideration. The difference in the air
density in altitude somehow offsets the presented asymmetry of the
cycle in regards to the GTs power rating. However, the mission profile of
each aircraft included the extra power demands in the case of one
engine out EOR safety case, whilst both the GTs and the energy storage
were slightly oversized in every proposed propulsion system.
The limitations of the Simulink models of the machines (i.e.
superconducting generators and cryo-cooler) have already been
mentioned. To these we should add the limitations of the battery sizing
model itself. The model inputs are varied manually to ensure the battery
power and energy capacity can reduce the power requirements by the
212
generators, reducing their size and cooling requirements. A more robust
way of varying these inputs is necessary in the future models. This could
facilitate the generation of results in a more time efficient way.
Furthermore, other outputs such as volume and life cycles should be
included in higher TRL studies.
6.5 Roadmap for Novel Flight Cycles Investigation
In the previous sections several novel hybrid configurations using energy
storage were investigated. The majority of hybrid/electric approach proposals in
aviation industry are based on conventional configurations at least regarding the
propulsion system design. However, it is fair to claim that Turbo-electric
Distributed Propulsion (TeDP) concept represents a disruptive technology that
requires the synergy of several subsystems and a more integrated approach
between the airframe and the propulsion system design procedures. Thus, the
aircraft design must be adapted to best capitalise on the potential benefits of
this new propulsion system. A simple modification of an existing aircraft design
with a novel propulsion system will most probably lead to pessimistic results
since no optimisation will be made.
The first step towards the full hybridisation of future aircraft should be the
optimisation of the main power sources. The GTs will no longer provide thrust
and their new role will be the production of electric power. Hence, new
optimised GT architectures need to be designed and in the case of novel cycles
their role could be even more specific. The concept behind these hybrid
configurations is to optimise each component for a specific function and/or
phase of the flight (Malkin and Pagonis, 2015b).
Apart from the traditional GTs option, other power units could be also
investigated in these configurations. In Figure 107 an example of a novel HEDP
architecture is being presented. The power system is illustrated by a main
power bus bar that is supplied by several prime movers such as gas-turbine
alternators (GTAs), a Secondary Power Unit (SPU) and energy storage. The
role of the SPU will vary depending on the range, power requirements and the
213
overall architecture being chosen. Different SPU options could be explored such
as:
Reciprocating engines
Hydrogen Fuel Cells
Tailored optimised GTs
The GTAs might be either of different sizes or identical and the merits of both
options need to be explored. A full performance assessment of all the SPU
options could affirm the ideal candidate for each aircraft.
Figure 107 HEDP Architecture Proposal
A full analysis of the flight envelope of each aircraft could identify the potential
dynamic requirements of their whole flight cycle. These dynamic requirements
could be handled by the energy storage mechanisms in order the rest of the
power units to perform more efficiently in a constant power rating. This was
basically investigated in the previous sections (6.2-6.3) and its feasibility in
terms of weight was confirmed in most of the cases. However, as it was already
pointed out a more extensive study of the energy storage mechanisms is
necessary since their design procedure is a complicated task that differs
depending on the application.
The dynamic analysis of the power requirements of the aircraft power network
should be one of the last stages of the novel flight cycles research study. In this
214
stage different rates of climb and descent could be investigated. Cruising
altitude and speed could also be altered to match the propulsion system
maximum efficiency. Note that power does not lapse with altitude in the case of
electrical machines and hence new criteria need to be met in case of the hybrid
aircraft. Finally, the already more electric approach will enhance the use of
electric power for phases such as taxiing and for secondary systems such as
the ECS, anti-icing, landing gear etc. The benefits of such an electric approach
should be clearly pointed out, quantified and combined with the propulsion
system hybridisation benefits.
In conclusion future studies of hybrid configurations should focus on the
following issues:
- reduce peak power demands of the main prime movers
-efficient management of dynamic requirements of the flight mission
-use of energy recovery at various stages
-varying operating cycles’ factors such as cruising altitude, speed, climb and
descent rate to further optimise each hybrid configuration
-possibility of electric taxiing without any extra equipment
-increase safety for certain failure modes
215
7 Conclusions and Future Work
The aerospace industry has been pressured to develop more environmental
friendly aircraft designs for the next generation commercial aircraft. Aggressive
targets in regards to emissions (i.e. noise, NOx, and fuel consumption) have
been set by both American and European institutions. Conventional
configurations seem unable to satisfy these optimistic goals; hence there is a
general interest in technologies which can be considered disruptive.
Distributed Propulsion (DP) technology is one of the most promising concepts
which could make a positive impact on environment in the following years.
There are several DP modifications which have been explored throughout the
years, but the most beneficial seems to be the Hybrid Electric Distributed
Propulsion (HEDP). This HEDP concept is nonetheless associated with various
supporting technologies, many of which are still in an early stage of
development.
Superconducting technology is considered the main enabler for the whole
concept due to the weight and efficiency improvements that could become
feasible by the implementation of a fully Superconducting Power Network (SPN)
in the propulsion system of such an aircraft. This network will include many
novel elements, some of which are still in an embryonic state. The concurrent
use of these elements in the same network creates several unknowns in the
system design which will require additional and extended experimental work to
be fully understood. Moreover, a SPN requires constant cooling to cryogenic
temperatures in order to perform according to its full potentials. Several studies
have been focused on the optimisation of the cooling system in a HEDP
configuration.
7.1 Concluding Remarks
In this project, several aspects of future HEDP configurations have been
investigated. First of all, the role of SPNs has been explored looking at the
constraints and the potential benefits that such a network could bring in a DP
design. The reduced weight of superconducting components is one of the
216
already well-established characteristics of these networks. In this study, a
method of estimating the weight of fully superconducting machines was
established and corresponding models were developed and used during the
DEAP project. These models were also used in the novel concept of
Superconducting Electric Aircraft (SEA) that was investigated in Chapter 5. SEA
is a modification of the More Electric Aircraft (MEA) which was proposed as an
enabling technology for the extension of MEA concept to aircraft of different
sizes. Finally, different HEDP conceptual configurations with enhanced use of
energy storage mechanisms were explored in the final chapter of this research
study focusing on their feasibility in terms of weight for future aircraft designs.
The following concluding remarks were derived from this study:
7.1.1 Superconducting Power Networks (SPNs)
It became clear in the early stages of this research study that it is impossible to
model and simulate the performance of a SPN by using conventional modelling
strategies and without any further experimental work. The true zero resistance
of superconducting DC networks complicates the modelling process, whilst the
current sharing in the transmission system of such a network cannot be
predicted. The design procedure of an autonomous power network is a
complicated task which has not be fully analysed in the recent literature. The
hybrid/electric ship is the most recent example of power network designed in a
similar way to the HEDP aircraft design. However, even this example does not
fully cover the particularities that a SPN brings in the design process. Starting
from the basic parameters selection a SPN follows completely different criteria
and priorities compared to a more conventional network. Different, novel
components for power generation, transmission, protection, and switching will
also be present in a superconducting configuration. The concurrent presence of
so many novel devices might complicate the performance prediction of such a
network but also adds flexibility in the design procedure. This flexibility has been
proven both in the SEA and in the HEDP concepts analysed in this research
study.
217
7.1.2 Superconducting Electrical Machines
The vast majority of superconducting machines currently being commercially
available have only their rotor primarily made by superconducting materials,
whilst a more conventional stator is normally used. However, in a HEDP
configuration fully superconducting machines will be required in order to fully
capitalise the weight and efficiency benefits of such a type of machines. A
Simulink model that can estimate the weight of these machines has been
developed for the purposes of this research study as well as for the purposes of
the DEAP project. Two different versions of this method have been designed
and compared with the NASA weight figures for superconducting machines.
The more conservative method could be considered as a pessimistic prediction
method for the weight estimation of these machines but its reliability has been
verified by the use of these models in the DEAP project weight calculations.
Furthermore, conventional environmental screens (i.e. iron and conducting)
proved to be either too heavy or too inefficient especially for high power
machines. On the other hand, novel superconducting screens might be the
solution for these fully superconducting machines.
7.1.3 Superconducting Electric Aircraft (SEA)
A novel concept aiming on further enhancing the already successful More
Electric Aircraft (MEA) concept was also proposed in this project. The feasibility
in terms of weight of using a superconducting network for the secondary power
systems of aircraft of different sizes was explored. Results show that SEA
aircraft becomes weight beneficial for electric power requirements over 1.5 MW,
whilst it is still weight neutral for electric loads around 1MW (Boeing 787 case).
The anticipated further electrification of future aircraft might enhance the
attractiveness of SEA concept even further. Besides, SEA designs will also be
more efficient, easily scalable, and more fault tolerant than the conventional
equivalent versions. Finally, whilst in a MEA approach the use of a SPN might
be considered as optional, in a HEDP type of aircraft its use appears to be
necessary. The latter is dictated by the over an order of magnitude higher
electric power requirements in the case of a HEDP system.
218
7.1.4 Novel Flight Cycles with Energy Storage
The flexibility of several HEDP configurations using energy storage
mechanisms was also investigated in this research study. The use of energy
storage as a main prime mover during take-off, as the main power source
during cruise, and as a boost supplementary power unit during take-off and
climb were considered for a number of different aircraft cases including the
DEAP baseline aircraft. Different battery and supercapacitor technologies were
considered as potential candidates for such configurations. The study showed
that using a battery pack as a boost unit during the demanding phases of take-
off and climb seems as the most promising configuration in terms of weight for
most of the aircraft sizes. On the other hand, using energy storage during cruise
showed some very pessimistic results in regards to the weight feasibility of such
a system. Lithium sulphur batteries proved to be the most weight efficient
energy storage option in almost all cases explored. Their low cyclability
however might give the advantage to future Lithium ion technologies especially
for shorter range aircraft cases. Similar to the SEA case, using energy storage
in an aircraft propulsion system could bring extra benefits in terms of efficiency,
redundancy, and flexibility of the whole aircraft power network.
7.1.5 Key Findings Summary
The key findings of this research study could be summarised as follows:
Lack of appropriate simulation tool for the steady-state models of a SPN
due to the uncertainty of the current splitting in these networks
Different basic parameters selection in the case of a SPN (higher normal
currents which will be easier to handle)
More than 7% increased overall network efficiency with a
superconducting transmission system
Novel Simulink models to estimate the weight of fully superconducting
machines based on the TRV method were developed
Requirement for a novel type of environmental screen for SMs has been
identified since the conventional types (i.e. iron and conducting) are too
heavy and/or inefficient
219
A superconducting version of a MEA Boeing 787 type of aircraft can be
almost two times lighter based on the current technology being used or
weight neutral if future technology of conventional equipment is used
In larger aircraft (i.e. A380 size) the SEA could be more than two times
lighter even compared to future technology predictions
SEA concept becomes weight beneficial when the aircraft electric load
demand is around 1.5 MW or higher
Lithium sulphur technology is the most promising technology in terms of
weight for the HEDP configurations investigated here
An all-electric aircraft proved to be significantly heavier than the rest of
the hybrid configurations at least based on the assumptions being made
in this study
The use of batteries as a supplementary power unit during demanding
flight phases showed the most promising results in terms of weight for
hybrid configurations
Batteries could also be used as the main power source during take-off in
short range aircraft without any weight penalty
7.2 Recommendations for future work
This project investigated several technologies that are still in an early stage of
development; hence there is a wide range of activities that can be
recommended as future work in this field. In the concluding remarks of each
chapter the technology gaps have been identified and suggestions for future
work have been pointed out. Some of the areas that seem more essential to be
explored will be prioritised in this sub-section.
- Extensive laboratory work in regards to the superconducting power
networks. It has been pointed out numerous times that
superconductivity is the main enabler of the disruptive concepts
presented in this research study and form the most promising concepts
for future aviation. However, there are several issues related to the use
220
of these networks. First of all, the uncertainty of their steady-state and
dynamic performance. Superconducting DC networks are characterised
by their zero resistance which might seem as an ideal case but at least
for these preliminary studies it also creates certain unknowns. The
current sharing on these networks cannot be predicted and experimental
work on this field is urgent. The resistive divider effect that typically splits
the current into the several nodes of the network is not present in a SPN.
Although early experimental studies (Pei et al., 2012) reported normal
current distribution for multi-strand MgB2 wires in an AC system, a
superconducting DC network load flow still remains a mystery.
Furthermore, superconducting equipment such as SFCLs should be
optimised for airborne applications. Also, the protection coordination of
these networks is a field that needs extra attention.
- Cooling system further studies are essential. The two cooling system
options which were presented in the literature section need to be
optimised for aerospace applications. Especially the cryo-coolers
currently used in the industry have not be designed in the most weight
efficient way. Furthermore, the models used for the cryo-coolers’ weight
estimation for the DEAP project are approaching the limit of their
capability. New models are required which will take into account the
difference in components efficiency as well as they will predict the mass
based on component-level estimations and not overall system
considerations (Palmer, Pagonis and Malkin, 2015). Concerning the
cryogenic fluid solution a whole different systems’ approach is
necessary. Issues such as location and volume of the tank, liquid
hydrogen production and storage in the airports, and advanced fuel
system development are just a few examples of the fields which need
additional extensive studies. The first two recommended future studies
are strongly related since the feasibility of SPNs highly depends on the
existence of an efficient and lightweight cooling system.
- Components’ performance in cryogenic temperatures needs to be
fully understood. Elements such as cryogenic power converters, fully
221
superconducting machines, and superconducting switching devices need
to be widely produced and used in less fault sensitive applications before
the 2035 timeframe which has been set as a goal for the next generation
aircraft. Furthermore, the interaction between all these novel components
needs to be observed in a real network situation.
- Energy storage more application-specific study is necessary. The
design of a battery system is a complicated task which is more
application specific than most people believe. The majority of preliminary
studies (including the current one) looking at energy storage devices
simply assume a power or energy density value to estimate their weight
without taking into account any other characteristics of these systems.
Especially when a future technology is being used for these studies,
overoptimistic predictions are typically being made and parameters such
as safety, life cycles and environmental impact are simply ignored.
222
223
224
REFERENCES
(CAA), C.A.A. (2002) AIRWORTHINESS INFORMATION LEAFLET CAA,
Abdel-Fadil, R., Eid, A. and Abdel-Salam, M. (2013) ‘Electrical Distribution Power Systems of Modern Civil Aircrafts’, 2nd International Conference on Energy Systems and Technologies, , pp. 201–210.
Abdelhafez, A.A. and Forsyth, A.J. (2009) ‘A review of More Electric Aircraft’, 13rd International Conference on Aerospace Sciences &Aviation Technology. Cairo, Egypt.
Adams, C. (2001) A380: ‘More Electric Aircraft’ Avionics Today.
Akimitsu, J. (2001) ‘Discovery of MgB2 Superconducting Material (unpublished)’, Symposium on Transition Metal Oxides.
Alderman, J. (2014) DEAP QMM3- Airbus Group Work Package 1 Summary.
Amatucci, G.G., Badway, F. and DuPasquier, A. (2000) ‘Novel Asymmetric Hybrid Cells and the Use of Pseudo-reference Electrodes in Three Electrode Cell Characterization’, in Nazri, G. A., Thackeray, M. M. and Ohzuku, T. (eds.) Intercalation Compounds for Battery Materials. , p. 344.
Ameyogo, G. (2007) Distributed Propulsion and Future Aerospace Technologies. Cranfield University.
Ameyogo, G. and Singh, R. (2007) ‘Advanced Cycles for Distributed Propulsion’, ISABE (ed.)18th International Symposium on Airbreathing Engines. Beijing.
Amy, J. (2005) ‘Modem, high-converter-populations argue for changing how to design naval electric power systems’, IEEE Electric Ship Technologies Symposium. Philadelphia, PA: IEEE.
Andersson, G. (2006) Modelling and Analysis of Electric Power Systems. Zurich.
Armstrong, M.J., Christine, A.H.R., Blackwelder, M.J. and Rajashekara, K. (2012) ‘Propulsion System Component Considerations for NASA N3-X Turboelectric Distributed Propulsion System’, SAE International J. Aerospace, 5(2), p. 344.
Ashcraft W., S., Padron S., A., Pascioni A., K., Stout W., G. and Huff L., D. (2011) ‘Review of Propulsion Technologies for N+3 Subsonic Vehicle Concepts’, NASA technical memorandum, 2011-21723
225
AVO (2015) Transient Motor Starting Analysis.
Ayre, J. (2014) Lithium-Air Batteries Take Step Forward EVObsession.
Baran, M. and Mahajan, N. (2007) ‘Overcurrent Protection on Voltage-Source-Converter- Based Multiterminal DC Distribution Systems’, IEEE Transactions on Power Delivery, 22(1), p. 406.
Barnes, P.N., Sumption, M.D. and Rhoads, G.L. (2005) ‘Review of high power density superconducting generators- Present state and prospects for incorporating YBCO windings’, Cryogenics, 45(10-11), pp. 670–686.
Barraud, G., Thomas, M., Charr, N., Nespoulous, S. and Smaoui, H. (2015) ‘Study of Hybrid Electric Propulsion for a Radical Aircraft Concept’, MEA 2015- More Electric Aircraft. Toulouse, France.
batteryuniversity.com (2015) BU-205: Types of Lithium Ion.
Bellis, M. (2015) History of Superconductors., about.com
Berg, F. and Dodds, G. (2013) PSAM Aerospace Applications Report. PSAM Project.
Berg, F., Palmer, J., Miller, P., Husband, M. and Dodds, G. (2015a) ‘HTS Electrical System for a Distributed Propulsion Aircraft’, IEEE Transactions on Applied Superconductivity, 25(3)
Berg, F., Palmer, J., Pagonis, M. and Malkin, D. (2015b) DEAP D2-2c Superconducting System Model. Unpublished: Airbus Group.
Biswas, A., Khan, M.E. and Sarker, U. (2013) ‘Transient Stability Improvement of a Conventional Power System by Superconducting Fault Current Limiter’, Global Journals of Researches in Engineering, 13(5), pp. 25–30.
Blaugher, R.D., Parker, J.H. and McCabria, J.L. (1977) ‘HIGH SPEED SUPERCONDUCTING GENERATOR’, IEEE Transactions on Magnetics, 13(1), p. 755.
Boeing (2013) 787 Dreamliner Electrical System.
Bollman, A.M., Armstrong, M.J., Jones, C.E., Norman, P.J. and Galloway, S.J. (2015) ‘Development of Voltage Standards for Turbo-electric Distributed Propulsion Aircraft Power Systems’, Electrical Systems for Aircraft, Railway, Ship Propulsion and Road Vehicles (ESARS), 2015 International Conference. Aachen: IEEE, p. 1.
226
Brown, G. V (2011) ‘Weights and Efficiencies of Electric Components of a Turboelectric Aircraft Propulsion System’, 49th AIAA/ASME/SAE/ASEE Joint Propulsion Conference. Orlando, Florida: AIAA.
Bumby, J.R. (1983) Superconducting rotating electrical machines (Monographs in Electrical and Electronic Engineering). Clarendon Press.
Buquet, M. (2007) Distributed Propulsion: An Overview of Possible Transmission Technologies. Cranfield University.
Chambon, T. (2012) Impact of the boundary layer ingestion on the noise generated by distributed propulsion concept. Cranfield University.
Chengyuan, L., Doulgeris, G., Laskaridis, P. and Singh, R. (2012) ‘Turboelectric Distributed Propulsion System Modelling for Hybrid-Wing-Body Aircraft’, 48th AIAA/ASME/SAE/ASEE Joint Propulsion Conference. Atlanta, Georgia: AIAA.
Christensen, K., Albertus, P., Sanchez-Carrera, R.S., Lohmann, T., Kozinsky, B., Liedtke, R., Ahmed, J. and Kojic, A. (2012) ‘A Critical Review of Li/Air Batteries’, Journal of Electrochemical Society, 159(2), p. 1.
Clark, S.F. (2012) 787 Propulsion System Boeing Aeromagazine.
Costi, F. (2012) Investigation on boundary layer ingestion effects on a ducted axial fan for a distributed propulsion system. Cranfield University.
Cotton, I., Nelms, A. and Husband, M. (2007) ‘Defining Safe Operating Voltages for Aerospace Electrical Systems’, Electrical Insulation Conference and Electrical Manufacturing Expo, 2007. Nashville, TN: IEEE, p. 67.
Dodson, B. (2013) New Lithium/sulphur battery doubles energy density of lithium-ion.
Doerry, N.H. and Fireman, H. (2006) ‘Designing All Electric Ships’, Proceedings of the Ninth International Marine Design Conference. Ann Arbor, MI, p. 475.
Dombrovski, V., Driscoll, D., Shoykhet, B.A., Umans, S.D. and Zevchek, J.K. (2005) ‘Design and Testing of a 1000-hp High-Temperature Superconducting Motor’, IEEE Transactions on Energy Conversion, 20(3), p. 638.
Doulgeris, G., Lelias, G., Laskaridis, P., Pilidis, P. and Singh, R. (2012) ‘Dynamic Response and High Cycle Fatigue Analysis of Fan Blades under Inlet Distortion’, 48th AIAA/ASME/SAE/ASEE Joint Propulsion Conference. Atlanta, Georgia.
Ebner, A. (2007) ‘Transient Transformer Inrush Currents due to Closing Time- and Residual Flux Measurement- Deviations if Controlled Switching is used’, Proceedings of European EMTP-ATP Conference. Leon, Spain.
227
Eckels, P.W. and Snitchler, G. (2008) ‘5 MW High Temperature Superconductor Ship Propulsion Motor Design and Test Results’, Electric Machines Technology Symposium, 117(4) Naval Engineers Journal, p. 31.
Edwards, H. (2011) Extract from DNS171343, Sugar volt (hybrid electric aircraft) assessment - The future of energy storage (20 year time frame). SRC, Derby, UK: Rolls Royce Private Data.
Eggenspieler, F. (2006) Distributed Propulsion Aircraft Wing Structure and Systems Design. Cranfield University.
Electronics, M. (2015) Maxwell 16V Ultracapacitor Modules .
Eriksson, J.T., Mikkonen, R., Paasi, J., Perala, R. and Soderlund, L. (1997) ‘A HTS Synchronous Motor at Different Operating Temperatures’, IEEE Transactions on Applied Superconductivity, 7(2), p. 523.
Felder, J.L., Brown, G. V, Kim, H.D. and Chu, J. (2011a) ‘Turboelectric Distributed Propulsion in a Hybrid Wing Body Aircraft’, 20th International Society for Airbreathing Engines (ISABE 2011). Gothenburg, Sweden.
Felder, J.L., Kim, H.D. and Brown, G. V (2009) ‘Turboelectric Distributed Propulsion Engine Cycle Analysis for Hybrid-Wing-Body Aircraft’, 47TH AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition. Orlando, Florida: AIAA.
Felder, J.L., Kim, H.D., Brown, G. V and Chu, J. (2011b) ‘An Examination of the Effect of Boundary Layer Ingestion on Turboelectric Distributed Propulsion Systems’, 49th AIAA/ASME/SAE/ASEE Joint Propulsion Conference. Orlando,Florida: AIAA.
Felder, J.L., Tong, M.T. and Chu, J. (2012) ‘Sensitivity of Mission Energy Consumption to Turboelectric Distributed Propulsion Design Assumptions on the N3-X Hybrid Wing Body Aircraft’, 48th AIAA/ASME/SAE/ASEE Joint Propulsion Conference. Atlanta, Georgia: AIAA.
Fletcher, S.D.A., Norman, P.J., Galloway, S.J. and Burt, G.M. (2011) ‘Determination of Protection System Requirements for DC UAV Electrical Power Networks for Enhanced Capability and Survivability’, IET Electrical Systems in Transportation Journal, 1(4), p. 137.
Furmanczyk, K. (2009) ‘Power Conversion Technologies for Improved System Performance’, 2009 Join Service Power Expo. New Orleans, Louisiana: Crane Aerospace and Electronics.
Gamble, B., Kalsi, S., Snitchler, G. and Madura, D. (2002) ‘The Status of HTS Motors’, IEEE (ed.)Power Engineering Society Summer Meeting. Chicago, IL, USA, Vol.1, p. 270.
228
Gamble, B., Snitchler, G. and MacDonald, T. (2011) ‘Full Power Test of a 36.5 MW HTS Propulsion Motor’, IEEE Transactions on Applied Superconductivity, 21(3), p. 1083.
Gibson, A.R., Hall, D.W., Schiltgen, B.T., Foster, T.J., Keith, J. and Masson, P. (2010) ‘The Potential and Challenge of TurboElectric Propulsion for Subsonic Transport Aircraft’, 47TH AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition. Orlando, Florida.
Gieras, J.F. (2009) ‘Superconducting Electrical Machines State of the Art’, Przeglad Elektrotechniczny, 85(12), pp. 1–19.
Glover, J.D., Sarma, M.S. and Overbye, T.J. (2010) Power System Analysis and Design.Fifth. Global Engineering: Christopher M. Shortt.
Gohardani, A.S., Doulgeris, G. and Singh, R. (2011) ‘Challenges of future aircraft propulsion: a review of distributed propulsion technology and its potential application for the all electric commercial aircraft’, Progress in Aerospace Sciences, 47(5), pp. 369–391.
Golovashkin, A.I., Ivanenko, O.M., Kudasov, Y.B., Mitsen, K. V., Pavlovsky, A.I., Platonov, V. V. and Tatsenko, O.M. (1991) ‘LOW TEMPERATURE DIRECT MEASUREMENTS OF Hcz IN HTSC USING MEGAGAUSS MAGNETIC FIELDS’, Physica C: Superconductivity, 185-189, pp. 1859–1860.
Graham R., W., Hall A., C. and Morales V., M. (2014) ‘The potential of future aircraft technology for noise and pollutant emissions reduction’, Transport Policy Special Issue Aviation and the Environment, 34, pp. 36–51.
Green, M.W., Schiltgen, B.T. and Gibson, A.R. (2012) ‘Analysis of a Distributed Hybrid Propulsion System with Conventional Electric Machines’, 48th AIAA/ASME/SAE/ASEE Joint Propulsion Conference. Atlanta, Georgia.
Griffith, A.A. (1954) Improvements relating to aircraft and aircraft engine installations.
Guynn D., M., Berton J., J., Hendricks S., E., Tong T., M., Haller J., W. and Thurman R., D. (2011) ‘Initial Assessment of Open Rotor Propulsion Applied to an Advanced Single-Aisle Aircraft’, 11th AIAA Aviation Technology, Integration, and Operations (ATIO) Conference. Virginia Beach, VA.
Guzik, M.C. and Tomsik, T.M. (2011) ‘An Active Broad Area Cooling Model of a Cryogenic Propellant Tank with a Single Stage Reverse Turbo- Brayton Cycle Cryocooler’, Thermal and Fluids Analysis Workshop., pp. 1–15.
Hale, J. (2008) Boeing 787: From the Ground Up Boeing Aeromagazine.
Hampton, K. (2013) Small but mighty Energy Storage Journal.
229
Harada, H. (2003) ‘High Temperature Materials for Gas Turbines: The Present and Future’, Proceedings of the International Gas Turbine Congress. Tokyo.
Harrop, P. (2013) Global marine electric vehicle market to reach $6.3bn in 2023 . IDTechEx Research report.
Hassannia, A. and Darabi, A. (2014) ‘Design and Performance Analysis of Superconducting Rim-Driven Synchronous Motors for Marine Propulsion’, IEEE Transactions on Applied Superconductivity, 24(1)
Hendershot, J.R. and Miller, T.J.E. (2010) Design of Brushless Permanent-Magnet Machines. Venice, Florida: Motor Design Books LLC.
Hennessy, M. (2009) Lightweight, Efficient Power Converters for Advanced Turboelctric Aircraft Power Systems MTECH Laboratories, LLC,
Hirst, M., McLoughlin, A., Norman, P.J. and Galloway, S.J. (2011) ‘Demonstrating the More Electric Engine: a step towards the power optimised aircraft’, IET Electrical Power Applications, 5(1), p. 3.
Hughes, A. and Miller, T.J.E. (1977) ‘Analysis of fields and inductances in air-cored and iron-cored machines’, Proceedings of the Institution of Electrical Engineers, 124(2), p. 121.
Imanishi, N. and Yamamoto, O. (2014) ‘Rechargeable lithium-air batteries: characteristics and prospects’, Materials Today, 17(1), p. 24.
Isikveren, A.T., Seitz, A., Vratny, P.C., Pornet, C., Plotner, K.O. and Hornung, M. (2012) ‘Conceptual Studies of Universally-Electric Systems Architectures Suitable for Transport Aircraft ’, 61. Deutscher Luft- und Raumfahrtkongress 2012 . Berlin, Germany.
Iwakuma, M., Tomioka, A., Konno, M., Hase, Y., Satou, T., Iijima, Y., Saitoh, T., Yamada, Y., Izumi, T. and Shiohara, Y. (2007) ‘Development of a 15 kW Motor With a Fixed YBCO Superconducting Field Winding’, IEEE Transactions on Applied Superconductivity, 17(2), p. 1607.
Jin, J.X. (2007) ‘High Efficiency DC Power Transmission Using High-Temperature Superconductors’, Physica C: Superconductivity, 460-462, p. 1443.
Jin, J.X., Xin, Y., Wang, Q.L., He, Y.S., Cai, C.B., Wang, Y.S., Wang, Z.M., Fault, A.H.T.S. and Limiters, C. (2014) ‘Enabling High-Temperature Superconducting Technologies Toward Practical Applications’, 24(5)
Jones, R.J. (2002) ‘The More Electric Aircraft - Assessing the Benefits’, Proceedings of the Institution of Mechanical Engineers, 216(G), p. 259.
230
Joshi, C.H., Prum, C.B., Schiferl, R.F. and Driscoll, D.I. (1995) ‘Demonstration of Two Synchronous Motors Using High Temperature Superconducting Field Coils’, IEEE Transactions on Applied Superconductivity, 5(2), p. 968.
Jow, T.R., Xu, K., Allen, J., Borodin, O., Cresce, A., Behl, W., Marx, M. and Allen, J. (2014) High Energy and High Power Li-ion Batteries.
Keison (2014) Keyson Products.
Kim, B., Chung, H. and Kim, W. (2012) ‘High-Performance Supercapacitors based on vertically aligned carbon nanotubes and nonaqueous electrolytes’, IOP Publishing Nanotechnology, 23, p. 1.
Kim, H., Felder, J., Tong, M. and Armstrong, M. (2013) ‘Revolutionary Aeropropulsion Concept for Sustainable Aviation: Turboelectric Distributed Propulsion’, ISABE Conference, , pp. 1–12.
Kim, H.D. (2010) ‘Distributed Propulsion Vehicles’, 27th International Congress of the Aeronautical Sciences. Nice, France: ICAS.
Kim, H.R., Sim, J. and Hyun, O.B. (2006) ‘Temperature distribution in SFCLs based on Au/YBCO films during quenches’, Cryogenics, 46(4), pp. 305–311.
Kim, S.-H. and Hahn, S.-Y. (2000) ‘Analysis and Design of a Induction Generator with a Superconducting Bulk Magnet Rotor’, IEEE Transactions on Applied Superconductivity, 10(1), p. 931.
Ko, A., Leifsson, L.T., Schetz, J. a, Mason, W.H. and Grossman, B. (2003) ‘MDO of a Blended-Wing-Body Transport Aircraft with Distributed Propulsion’, AIAA ’ s 3rd Annual Aviation Technology , Integration , and Operations ( ATIO ) Technical Forum. Denver , CO.
Kongsberg (2015) Marine switchboards., http://www.km.kongsberg.com/ks/web/nokbg0240.nsf/AllWeb/6591A4733274161DC125766400373AAF?OpenDocument
Kwon, Y.K., Sohn, M.H., Baik, S.K., Lee, E.Y., Kim, J.M., Monn, T.S., Park, H.J., Kim, Y.C. and Ryu, K.S. (2005) ‘Development of a 100 hp Synchronous Motor With HTS Field Coils’, IEEE Transactions on Applied Superconductivity, 15(2), p. 2194.
Larbalestier, D., Gurevich, a, Feldmann, D.M. and Polyanskii, a (2001) ‘High-Tc superconducting materials for electric power applications.’, Nature, 414(6861), pp. 368–377.
Laskaridis, P. and Pilidis, P. (2004) Performance investigations and systems architectures for the more-electric aircraft. Cranfield University.
231
Laskaridis, P., Valencia, E., Kirner, R. and Wei, T.J. (2015) ‘Assessment of Distributed Propulsion Systems Used with Different Aircraft Configurations’, 51st AIAA/SAE/ASEE Joint Propulsion Conference. Orlando, Florida.
Leong, K.K. (2011) Utilising Power Devices Below 100 K to Achieve Ultra-low Power Losses. http://wrap.warwick.ac.uk/46807/: University of Warwick.
Li, G.Z., Sumption, M.D., Susnrer, M.A., Yang, Y., Reddy, K.M. and Rindfleisch, M.A. (2012) ‘The Critical Current Density of Advanced Internal-Mg-Diffusion-Processed MgB2 Wire’, IOP Publishing, (Superconductor Science and Technology)
Lieberman, M.A. and Lichtenberg, A.J. (2005) Principles of plasma discharges and materials processing.Second (ed.) Hoboken, N.J.: Wiley-Interscience.
Liu, C. (2013) Turboelectric Distributed Propulsion System Modelling. Cranfield University.
Luongo, C.A., Masson, P.J., Nam, T., Mavris, D., Kim, H.D., Brown, G. V, Waters, M. and Hall, D.W. (2009) ‘Next Generation More-Electric-Aircraft: A Potential Application for HTS Superconductors’, Applied Superconductivity, 19(3), pp. 1055–1068.
Maeda, H., Tanaka, Y., Fukutomi, M. and Asano, T. (1988) ‘A new high-T (c) oxide superconductor without a rare earth element’, Japanese Journal of Applied Physics, 27(2), p. 209.
Maguire, J.F., Schmidt, F., Bratt, S., Welsh, T.E., Yuan, J., Allais, A. and Hamber, F. (2007) ‘Development and Demonstration of a HTS Power Cable to Operate in the Long Island Power Authority Transmission Grid’, Applied Superconductivity, 17(2), pp. 2034–2037.
Malkin, D. (2014) Superconducting Cables for Aerospace Electrical Systems. University of Manchester: Faculty of Engineering and Physical Sciences.
Malkin, P. and Pagonis, M. (2013) ‘The Design of Fully Superconducting Power Networks for Future Aircraft Propulsion’, 49th AIAA/ASME/SAE/ASEE Joint Propulsion Conference. San Jose, CA.
Malkin, P. and Pagonis, M. (2014) ‘Superconducting Electric Power Systems for Hybrid Electric Aircraft’, Aircraft Engineering and Aerospace Technology: An International Journal, 86(6), p. 515.
Malkin, P. and Pagonis, M. (2015a) ‘Future Energy Solutions For Hybrid-Electric Powered Aircraft’, The 22nd International Symposium on Air Breathing Engines. Phoenix, AZ USA, Vol.to be publ.
232
Malkin, P. and Pagonis, M. (2015b) ‘Propulsion System Optimisation Approach for Hybrid / Electric Aircraft’, More Electric Aircraft (MEA) Conference. Toulouse, France.
Manzel, C. (1924) Propelling mechanism for airships and the like.
marine.man.eu (2015) Diesel-electric Propulsion Plants.
Masson, P.J., Brown, G. V, Soban, D.S. and Luongo, C.A. (2007) ‘HTS machines as enabling technology for all-electric airborne vehicles’, Superconductor Science and Technology, 20(8)
Masuda, T., Yumura, H., Watanabe, M., Takigawa, H., Ashibe, Y., Suzawa, C., Kato, T., Okura, K., Yamada, Y., Hirose, M., Yatsuka, K., Sato, K. and Isojima, S. (2004) High-temperature Superconducting Cable Technology and Development Trends. SEI Technical Review.
Matsuzaki, H., Kimura, Y., Ohtani, I., Izumi, M., Ida, T., Akita, Y., Sugimoto, H., Miki, M. and Kitano, M. (2005) ‘An Axial Gap-Type HTS Bulk Synchronous Motor Excited by Pulsed-Field Magnetization With Vortex-Type Armature Copper Windings’, IEEE Transactions on Applied Superconductivity, 15(2), p. 2222.
McConnell, B.W., Walker, M.S. and Mehta, S.P. (2000) ‘HTS Transformers’, IEEE Power Eng. Rev., 20, pp. 7–11.
Meier, N. (2005) Civil Turboshaft/Turboprop Specifications.
Mikhaylik, Y., Kovalev, I., Schock, R., Kumaresan, K., Xu, J. and Affinito, J. (2015) High Energy Rechargeable Li-S Cells for EV Application. Status, Challenges and Solutions., Sion Power Corporation
Miller, T.J.E. (1989) Brushless Permanent Magnet and Reluctance Motor Drives. United States: Oxford University Press.
Miller, T.J.E. and Hughes, A. (1977) ‘Comparative design and performance analysis of air-cored and iron-cored synchronous machines’, Electrical Engineers, Proceedings of the Institution of Electrical Engineers, 124(2), p. 127.
Millikin, M. (2010) QinetiQ Zelphyr Used Sion Li-S Batteries in Record Unmanned Flight greencarcongress.com,
Millikin, M. (2009) FDK To Begin Mass Production of High-Capacity Li-Ion Capacitors; Automotive and Renewable Energy Applications
Moir, I. and Seabridge, A. (2013) Design and Development of Aircraft Systems. John Wiley & Sons Ltd.
233
Molina, M.G. (2010) ‘Dynamic Modelling and Control Design of Advanced Energy Storage for Power System Applications ’, in Brito, A. (ed.) Dynamic Modelling.
Montgomery, S. (2012) ElectroniCast- Solar PV Outdoor LED Lighting Using LIC
Murphy, T. (2012) The battery performance deficit disorder GIGAOM Research.
Nagata, H. and Chikusa, Y. (2014) ‘A lithium sulphur battery with high power density’, Journal of Power Sources, 264, p. 206.
Nakamura, T., Miyake, H., Ogama, Y., Motita, G., Muta, I. and Hoshino, T. (2006) ‘Fabrication and Characteristics of HTS Induction Motor by the Use of Bi-2223/Ag Squirrel-Cage Rotor’, IEEE Transactions on Applied Superconductivity, 16(2), p. 1469.
Van Noorden, R. (2014) ‘The rechargeable revolution: A better battery’, NATURE, 507(7490), p. 26.
Ojha, P. and Raghava, K. (2014) TRENT 1000: Aircraft Engine www.slideshare.com.
Oswald, B., Best, K.-J., Setzer, M., Soll, M., Gawalek, W., Gutt, A., Kovalev, L., Krabbes, G., Fisher, L. and Freyhardt, H.C. (2005) ‘Reluctance Motors with Bulk HTS Materials’, Superconducting Science and Technology, 18, p. 24.
Owano, N. (2012) Sugar Volt: Boeing puts vision to work in hybrid electric aircraft Phys.org.
Palmer, J., Pagonis, M. and Malkin, P. (2015) ‘Configuration Options for High-Power, Low Weight Aerospace Superconducting Distributed Propulsion Cryocoolers’, ISABE 2015 (to be published).
Palmer, J. and Shehab, E. (2015) ‘Modelling of Cryogenic Cooling System Design Concepts for Superconducting Aircraft Propulsion’, IET Electrical Systems in Transportation Journal, Pending pu
Palmer, J., Shehab, E., Fan, I.S. and Husband, M. (2013) ‘Quality Function Deployment and Sensitivity Analysis of Requirements for Future Aircraft Propulsion Cryogenic Cooling Systems’, Proceedings of the 11th International Conference on Manufacturing Research.
Paschen, F. (1889) Breakdown Voltage vs. Pressure Paschen’s Law and the ‘Paschen Curve’.
234
Pei, X., Smith, A.C., Husband, M. and Rindfleisch, M. (2012) ‘Experimental tests on a superconducting fault current limiter using three-strand MgB2 wire’, IEEE Transactions on Applied Superconductivity, 22(3)
Pei, X., Zeng, X., Smith, A.C. and Malkin, D. (2015) ‘Resistive Superconducting Fault Current Limiter Coil Design Using Multistrand MgB 2 Wire’, IEEE Transactions on Applied Superconductivity, 25(3)
Pilidis, P. (2012) Gas Turbine Theory and Performance
PolyPlus, B.C. (2009) Advanced Lithium Battery Technology.
Provost, M.J. (2002) ‘The More Electric Aero-Engine: A General Overview from an Engine Manufacturer’, Plc., R.-R. (ed.)Conference on Power Electronics, Machines and Drives. Derby: IET, p. 246.
Rada, A., Lubin, T., Lévêque, J., Berger, K., Douine, B., Rezzoug, A., Mezani, S., Didier, G., Hinaje, M. and Netter, D. (2015) ‘Study of a superconducting motor with high specific torque’, MEA 2015 More Electric Aircraft. Toulouse, France, p. 346.
Radebaugh, R. (2012) ‘Cryocoolers for aircraft superconducting generators and motors’, AIP Conf. Proc. ADVANCES IN CRYOGENIC ENGINEERING: Transactions of the Cryogenic Engineering Conference. Washington, USAs: AIP.
Ranade, S.J. and Xu, W. (1998) ‘Tutorial on Harmonics Modeling and Simulation (Chapter 1)’, Proc. IEEE Power Eng. Society, , pp. 1–7.
Rogers, S. (2012) Electric Drive Status and Challenges Energy Efficiency & Renewable Energy. U.S. Department of Energy,
Rosenkranz, C.A., Kohler, U. and Liska, J.L. (2007) ‘Modern Battery Systems for Plug-In Hybrid Electric Vehicles’, 23rd International Battery and Fuel Cell Electric Vehicle Symposium and Exposition (EVS 23). Anaheim, CA, USA.
Rosero, J.A., Ortega, J.A., Aldabas, E. and Romeral, L. (2007) ‘Moving towards a More Electric Aircraft’, IEEE Aerospace and Electronic Systems Magazine, 22(3), p. 3.
Ross, C., Armstrong, M., Blackwelder, M., Jones, C., Norman, P. and Fletcher, S. (2014) ‘Turboelectric Distributed Propulsion Protection System Design Trades’, SAE International, 2014-01-21
Roveglia, C. (2015) STABLE (STable high-capacity lithium-Air Batteries with Long cycle life for Electric cars.
235
Scanlan, R.M., Malozemoff, A.P. and Larbalestier, D.C. (2004) ‘Superconducting materials for large scale applications’, Proceedings of the IEEE, 92(10), pp. 1639–1654.
Schaub, U., Vlasic, E. and Moustapha, S. (1993) ‘Effect of Tip Clearance on the Performance of a Highly Loaded Turbine Stage’, Proceedings of AGARD PEP 82nd Symposium.
Schiltgen, B., Green, M.W., Gibson, A.R., Hall, D.W., Cummings, D.B. and Hange, C. (2012) ‘Benefits and Concerns of Hybrid Electric Distributed Propulsion with Conventional Electric Machines’, 48th AIAA/ASME/SAE/ASEE Joint Propulsion Conference. Atlanta, Georgia: AIAA.
Schiltgen, B.T., Green, M.W., Gibson, A.R., Foster, T.J. and Waters, M. (2011) ‘Split-Wing Propulsor Design and Analysis for Electric Distributed Propulsion’, 49th AIAA/ASME/SAE/ASEE Joint Propulsion Conference. Orlando, Florida: AIAA.
Scocco, D. (2006) S-Curves., Innovation Theory
Scott, A. (2014) New Boeing Jets hold key to more than half of future sales, Reuters, Aerospace May.
Senkov, O.N., Bhat, R.B. and Senkova, S. V (2004) ‘High Strength Aluminium Alloys For Cryogenic Applications’, Metallic Materials with High Structural Efficiency, , p. 151.
Shankland, S. (2014) Want major new aircraft designs? Wait until 2030., Cnet
Shcherbakov, A.V. (2011) Magnesium diboride superconductor : thermal stabilization and doping. University of Wollongong.
Shen, Y., Sun, D., Yu, L., Zhang, W., Shang, Y., Tang, H., Wu, J., Cao, A. and Huang, Y. (2013) ‘A high-capacity lithium-air battery with Pd modified carbon nanotube sponge cathode working in regular air’, Carbon, 62, p. 288.
Shi, J., Tang, Y., Zhou, Y., Chen, J., Xu, D., Wang, H., Lu, Y., Ren, L., Wei, B., Li, J. and Cheng, S. (2007) ‘Development of a conduction-cooled HTS SMES’, IEEE Transactions on Applied Superconductivity, 17(3), pp. 3846–3851.
Shuli, L. and Zhan, L. (2015) ‘Lithium-Sulphur Batteries’, in Rechargeable Batteries. Springer, p. 587.
Sim, J., Lee, K., Cha, G. and Lee, J.-K. (2004) ‘Development of a HTS Squirrel Cage Induction Motor With HTS Rotor Bars’, IEEE Transactions on Applied Superconductivity, 14(2), p. 916.
236
Simon, P. and Gogotsi, Y. (2010) ‘Charge storage mechanism in nanoporous carbons and its consequence for electrical double layer capacitors’, Philosophical Transactions of the Royal Society, 368, p. 3457.
Simon, P. and Gogotsi, Y. (2008) ‘Materials for electrochemical capacitors’, NATURE Materials, 7, p. 845.
Singh, T. (2013) EADS E-Thrust Electric Airliner Could Transform Air Travel by 2050 Inhabitat.
Sinnet, M. (2008) 787 No-Bleed Systems: Saving Fuel and Enhancing Operational Efficiencies Boeing Aeromagazine.
Skaar, S.E., Krovel, O. and Nilssen, R. (2006) Distribution, coil-span and winding factors for PM machines with concentrated windings NTNU Website.
Solovyov, V.F. and Li, Q. (2013) ‘Fast high-temperature superconductor switch for high current applications’, Applied Physics Letters, 103(3), p. 032603.
Sözeri, H., Özkan, H. and Ghazanfari, N. (2007) ‘Properties of YBCO superconductors prepared by ammonium nitrate melt and solid-state reaction methods’, Journal of Alloys and Compounds, 428(1-2), pp. 1–7.
Stephenson, D. (2010) Envisioning tomorrow’s aircraft www.boeing.com,
Stuckl, S., Van Toor, J. and Lobentanzer, H. (2012) ‘Voltair - The All Electric Propulsion Concept Platform - A Vision for Atmospheric Friendly Flight’, 28th Congress of the International Council of the Aeronautical Sciences . Brisbane, Australia.
Sutanto, D. and Cheng, K.W.E. (2009) ‘Superconductic Magnetic Energy Storage Systems for Power System Applications’, Proceedings of 2009 IEEE International Conference on Applied Superconductivity and Electromagnetic Devices. Chengdu, China, p. 377.
Takeda, T., Oota, T. and Togawa, H. (2006) ‘Development of Liquid Nitrogen-cooled Full Superconducting Motor’, Journal of Interactive Media In Education, 41(6)
THALES (2015a) Electrical Systems: Power Thinking.
THALES (2015b) Electrical Power Conversion Solutions- HVDC.
THALES (2015c) Electrical Power Conversion Solutions- AC to AC.
THALES (2015d) Advanced Electrical Power Conversion Solutions- Active Solutions.
237
Thiétart, R.A. (2013) The Boeing 787. The Dreamliner wakes up to reality ESSEC Business School Webpage.
Thome, R.J., Creedon, W., Reed, M., Bowles, E. and Schaubel, K. (2002) ‘Homopolar Motor Technology Development’, IEEE (ed.)Power Engineering Society Summer Meeting. Chicago, IL, USA, Vol.1, p. 260.
Tixador, P. and Daffix, H. (1997) ‘Conceptual design of an electrical machine with both low and high Tc superconductors’, IEEE Transactions on Applied Superconductivity, 7(4), p. 3858.
Torenbeek, E. (1992) Development and application of a comprehensive, design-sensitive weight prediction method for wing structures of transport category aircraft. Delft University of Technology.
Valencia, E. and Nalianda, D. (2015) ‘Methodology to assess the performance of an aircraft concept with distributed propulsion and boundary layer ingestion using a parametric approach’, Proceedings of the IMechE, 229(May), pp. 682–693.
Waltman, D.J. and Superczynski, M.J. (1995) ‘High-Temperature Superconducting Magnet Motor Demonstration’, IEEE Transactions on Applied Superconductivity, 5(4)
Warwick, G. (2013) eConcept - EADS’s Hybrid-Electric Airliner., Aviation Week Network (Things with Wings)
Wen, H., Bailey, W., Goddard, K., Al-Mosawi, M., Beduz, C. and Yang, Y. (2009) ‘Performance Test of a 100 kW HTS Generator Operating at 67 K–77 K’, IEEE Transactions on Applied Superconductivity, 19(3), p. 1652.
Whitlow B., J. and Sievers K., G. (1988) ‘NASA Advanced Turboprop Research and Concept Validation Program’, Conference and Exposition on Future Transportation Technology sponsored by the Society of Automotive Engineers. San Francisco, California: NASA Technical Memorandum 100891.
Whyatt, G.A. and Chick, L.A. (2012) Electrical Generation for More-Electric Aircraft using Solid Oxide Fuel Cells. Pacific Northwest National Laboratory Richland, Washington 99352: U.S. Department of Energy.
wikipedia.com (2015) Nissan Leaf.
Williams, M., Muley, A., Bolla, J. and Strumpf, H. (2008) ‘Advanced Heat Exchanger Technology for Aerospace Applications’, Seattle, Washington, United States.
238
Wolfgang, N., Grundmann, J. and Frauenhofer, J. (2012) ‘Test results from Siemens low-speed, high-torque HTS machine and description of further steps towards commercialisation of HTS machines’, Physica C, 482, p. 105.
Wright, J., Alderman, J., Berg, F., Raffaelli, L., Miller, P., Malkin, P. and Pagone, E. (2015) DEAP Final Technical Report
www.airlines-inform.com (2012) Commercial Aircraft of the World.
www.AviationExplorer.com (2015) Airbus A350 XWB Aircraft Airliner History Pictures and Facts.
www.batteryspace.com (2015) Category: LiFePO4 Batteries.
www.eaa-phev.org (2015) Toyota Prius Battery Specs.
www.what-when-how.com (2015) Electrical Conduction in Metals and Alloys (Electrical Properties of Materials) Part 2.
Xi, H.X., Gong, W.Z., Zhang, Y., Bi, Y.F., Ding, H.K., Wen Hou, B., H. and Xin, Y. (2006) ‘China’s 33.5m, 35kV/2 kA HTS ac power cable's operation in power grid’, Physica C: Superconductivity, 445-448, p. 1054.
Xin, Y., Han, Z. and Liao, Z. (2006) ‘Experimental 35 kV/21 MVA Superconducting Cable System Installed at PUJI Substation in China Southern Power Grid’, IEEE Trans. on Electrical and Electronic Engineering, 1(1), p. 8.
Yuan, W. (2011) Second-Generation High -Temperature Superconducting Coils and Their Applications for Energy Storage. Cambridge University.
Yuan, W., Xian, W., Ainslie, M., Hong, Z., Yan, Y., Pei, R., Jiang, Y. and Coombs, T.A. (2010) ‘Design and Test of a Superconducting Magnetic Energy Storage (SMES) Coil’, IEEE Transactions on Applied Superconductivity, 20(3), pp. 1379–1382.
Zhang, Y., Lehner, T., Fukushima, T., Sakamoto, H. and Hazelton, D. (2014) ‘Progress in production and performance of second generation (2G) HTS wire for practical applications’, 2013 IEEE International Conference on Applied Superconductivity and Electromagnetic Devices, 24(5), pp. 557–558.
239
APPENDICES
A.1 Conventional Electrical Machines Rotor and Stator
Dimensions
The following table includes the conventional machines that have been used in
4.2.2 in order to find a reliable relationship between the stator and rotor
diameter of electrical machines. The table consists of the required dimensions,
the type of the machines as well as their development stage at the moment
when these parameters were acquired. A further analysis of these machines is
not in the scope of this research study. All these data were collected by a Rolls
Royce internal study and the summary datasheet was available during the
DEAP project.
Table_A-1 Conventional Electrical machines dimensions
Machine type Development stage
Rotor Diameter
(mm)
Stator Diameter
(mm)
PM BLAC Prototype manufactured
66 144
PM BLAC Prototype manufactured
394 556
PM BLAC Prototype manufactured
265 390
SR Prototype manufactured
164 236
PM/VR Hybrid Prototype manufactured
99 160
PM BLAC Prototype manufactured
124 210
IM High level concept study
82.1 150
PM BLAC In production (Toyota Prius traction motor)
160.4 269.24
PM BLAC Detailed Concept Design
2450 2780
WF Synchronous In production (Trent 170 237
240
1000)
PM BLAC Detailed Concept Study
1040 1000
IM Prototype Manufactured
840 1267
PM BLAC Prototype Manufactured
116 160
A.2 Reverse Brayton Cryo-coolers (RBC) Simulink Model
A.2.1 Introduction and Simulink model overview
For project DEAP two different cryo-cooler models were developed by Joseph
Palmer (PhD student of Cranfield University, funded by RR), a single stage and
a double stage reverse-Brayton cryo-cooler. For the purposes of this study the
double stage version was used. Both systems are similar to operation with the
key difference that in the double stage case the heat rejection from the first
stage is removed by the second stage. Hence, the temperature difference
between the hot and cold parts is reduced enhancing the efficiency of the whole
system. There is an offset between the adding weight of the second stage and
the efficiency and potential reduction in input power of this type of cryo-coolers.
However, this version was preferred both because of the conservative weight
prediction and also due to its attractiveness in terms of efficiency. The following
figure illustrates a schematic diagram of a double stage Reverse-Brayton cryo-
cooler opposed to the actual overview of the Simulink model being developed.
241
Figure_A-1 Schematic diagram showing RBC next to the Simulink model
The main inputs and outputs of this model can be found in the following table.
The models are relatively complex based on numerous equations which are not
in the scope of this research study to be analysed. However, these models were
verified during the DEAP project by several experts and they were used for the
sensitivity studies carried out throughout the project; hence they could be
considered as reliable.
The cryo-cooler mass is calculated using equation (2-2), whilst for the two-stage
cryo-coolers of this study, this equation is used for each stage and the individual
mass values were added together. The compressor and turbine polytropic
efficiencies were assumed to be 0.90 and 0.92 respectively, assumptions
relatively conservative for the 2035 timeframe. The cold operating temperature
and the heat exchanger temperature delta could be varied depending on the
superconducting material and the coolant being used in each case. A pressure
drop of 5% was assumed for every heat exchanger of this model, whilst a heat
sink latent heat of 510 𝑘𝐽/𝑘𝑔 was considered.
242
Table_A-2 Main inputs/outputs of Cryo-cooler Simulink models
Inputs Units Outputs Units
Cold thermal load 𝑘𝑊 Heat sink required mass
flow rate
𝑘𝑔/𝑠
Cold operating temperature 𝐾 Required input power 𝑘𝑊
Heat Exchanger pressure drop - Cryo-cooler mass 𝑘𝑔
Heat exchange latent heat 𝑘𝐽/𝑘𝑔
Heat Exchanger Temperature
delta
𝐾
Compressor polytropic efficiency −
Turbine polytropic efficiency −
A.2.2 Main Assumptions and model limitations
There were a number of assumptions necessary to be made to reduce the
complexity and the number of variables of these models. These can be
summarised as follows:
There is no external heat transfer to the system (i.e. close system)
Transport losses were ignored (i.e. perfect transfer between components
was assumed)
Constant specific heat capacity values
Superconducting motors were used (models presented in Chapter 4.2)
Numerical assumptions taken based on current aerospace examples
available
The intermediate heat exchanger is defined by firstly taking the overall
temperature delta between the cold desired temperature and the heat
sink helium output temperature. This ratio is then square rooted. This
243
effectively implies the temperature ratio is equally shared among both
stages, allowing each system to be optimally efficient.
There are two main model limitations which need to be pointed out. First of
all, this is a steady-state model where no transient effects are taken into
consideration. Secondly, there are several technology uncertainties which
needed a number of assumptions to be made (i.e. heat exchanger
performance, components efficiency etc.). These values have not yet been
verified neither by experimental work nor by numerical methods and hence
the uncertainty in the estimation of the cryo-cooler mass remains high.
A.3 Turboshaft/Turboprop engines datasheet
In order to estimate the weight of the turboshaft engines for the Chapter 6
sensitivity study the specifications of several civil turboshaft/turboprop engines
were used. The following table summarises some of the engines which were
included in this study. Overall, 61 engines were included in the survey, but only
the most representative examples of each manufacturer were included in the
table that follows. The table also includes the application in which these engines
were used, the rated power and the overall dry weight in kg for each case. Note
that some of the reference engines were used in more than one application but
for reasons of space economy only one application per engine is mentioned.
244
Table_A-3 List of turboshaft/turboprop engines
Manufacturer Application Power (kW) Weight
(kg)
Allison Cesna 402/414 313.19 89.8
Allison (Rolls-Royce) MD600N 484.7 124.2
Avco Lycoming Cessna/Riley 421 458.6 142.88
Baranov (OMSK) An-38-200 1029 284.86
Garrett Commander 840/900 733.7 172.4
Garrett (Allied-Signal) Jetstream 4101 1284.8 281.2
General Electric Bell 214ST 1211.8 200.5
Isotov (Klimov) Mi-2/-2B/-2R 293.8 139.3
LHTEC Ayres LM200 2013.4 517.1
Mitsubishi MH-2000 653.2 154.2
Pratt Whitney Canada Starship 2000 894.8 229.5
Pratt Whitney Canada King Air F90 559.3 154.2
PZL Rzeszów W-3 Sokol 662.2 140.6
Rolls Royce Westland 30 Series 100 846.4 183.2
Saturn An-38 970.9 239.9
Turbomeca SA365C 477.9 120.2
Walter Ae-270 580.1 201.8
A.4 Aircraft Mission Profile in MATLAB
For the sensitivity study of Chapter 6 a MATLAB function has been developed
to produce the mission profile of each aircraft under investigation. This code
was based on work being produced during the DEAP project by the AGI but it
245
has been modified accordingly to match the requirements of the different aircraft
of the current study. The maximum power of the engines and the
battery/supercapacitor system was given as inputs combined with the maximum
energy storage energy available in each case. After that, the exact time duration
and power requirements of each phase are estimated and are given as outputs
from this MATLAB function and used as an input in the Simulink models
presented in Chapter 6.3. The following code was used for the Boeing 787
case and it is showed just as a representative example of a mission profile
MATLAB function.
function [ ] = missionmaker( )
%787 MISSIONMAKER Primes the workspace for the Chapter 6 study
Engine_1_max = 4200; %kW
Engine_2_max = 4200; %kW
Batt_P_max = 12500; %kW
Batt_P_min = 0; %kW
Batt_E_max = 111000; %kJ
assignin('base','Engine_1_max', Engine_1_max);
assignin('base','Engine_2_max', Engine_2_max);
assignin('base','Batt_P_max', Batt_P_max);
assignin('base','Batt_P_min', Batt_P_min);
assignin('base','Batt_E_max', Batt_E_max);
assignin('base','Batt_Chg', Batt_Chg);
% Power in kW, time in minutes
SLS_power = 10428.*2;
SLS_time = 1;
EoR_power = 10379.*2;
246
EoR_time = 8;
Climb_power = 4617.*2;
Climb_time = 20;
Cruise_power = 4049.*2;
Cruise_time = 915;
Descent_power = Cruise_power .* 0.8;
Descent_time = 20;
Landing_power = Cruise_power .* 1.4;
Landing_time = 2;
Taxi_power = Cruise_power .* 0.4;
Taxi_time = 5; %each end
mission = zeros(16,2);
mission(1,1) = 0;
mission(1,2) = Taxi_power;
mission(2,1) = Taxi_time .* 60;
mission(2,2) = Taxi_power;
mission(3,1) = mission(2,1) + 1;
mission(3,2) = SLS_power;
mission(4,1) = mission(3,1) + (SLS_time .* 60);
mission(4,2) = SLS_power;
mission(5,1) = mission(4,1) + 1;
mission(5,2) = EoR_power;
mission(6,1) = mission(5,1) + (EoR_time .* 60);
mission(6,2) = EoR_power;
247
mission(7,1) = mission(6,1) + 1;
mission(7,2) = Climb_power;
mission(8,1) = mission(7,1) + (Climb_time .* 60);
mission(8,2) = Climb_power;
mission(9,1) = mission(8,1) + 1;
mission(9,2) = Cruise_power;
mission(10,1) = mission(9,1) + (Cruise_time .* 60);
mission(10,2) = Cruise_power;
mission(11,1) = mission(10,1) + 1;
mission(11,2) = Descent_power;
mission(12,1) = mission(11,1) + (Descent_time .* 60);
mission(12,2) = Descent_power;
mission(13,1) = mission(12,1) + 1;
mission(13,2) = Landing_power;
mission(14,1) = mission(13,1) + (Landing_time .* 60);
mission(14,2) = Landing_power;
mission(15,1) = mission(14,1) + 1;
mission(15,2) = Taxi_power;
mission(16,1) = mission(15,1) + (Taxi_time .* 60);
mission(16,2) = Taxi_power;
%done = mission;
assignin('base','mission', mission);
end
248